Connecting to Azure Cosmos DB

For information on the module to import, and how to configure the necessary connection properties in a connection string, see Establishing a Connection.

Other available connection properties can be used to configure other aspects of the connector capabilities.

Python Version Support

Azure Cosmos DB Version Support

See Also

• Using the Connector: Establish connections and query Azure Cosmos DB through Python code.
• From SQLAlchemy: Use SQLAlchemy to establish a connection with dialect URL, and interact with Azure Cosmos DB data using mapped classes and Sessions.

Before You Connect

You can either create your own custom role definitions, or assign one of the built-in role definitions:

• CosmosDB Built-in Data Reader
• CosmosDB Built-in Data Contributor

You must also set the scope of the role assignment, where "/" means that the identity has access to all the databases.

For details, see Configure role-based access control for your Azure Cosmos DB account with Azure AD.

Dependencies

Installation

Windows:

pip install cdata_cosmosdb_connector-26.0.9655-cp310-abi3-win_amd64.whl

Linux:

pip install cdata_cosmosdb_connector-26.0.9655-cp310-abi3-linux_x86_64.whl

macOS:

pip install cdata_cosmosdb_connector-26.0.9655-cp310-abi3-macosx_12_0_arm64.whl

The macOS wheel supports arm64 (Apple Silicon) architectures only on macOS 12 and newer.

Regardless of the environment, certain distributions might require that the "pip3 install" command be used instead, to differentiate from a Python 2 distribution that might exist already. After installation, confirm whether the connector is successfully installed by running the "pip list" command. If "cdata_cosmosdb_connector" is present in the list output by the command, then the installation was successful.

Upgrading

When upgrading, "pip install" does not automatically clean up old JRE files. To avoid leftover files that could cause JVM errors, uninstall the previous version before installing the new one.

Licensing

After the installation is complete, a separate step is needed to activate a license for the connector. Among the CData assets in the distribution's site packages, there is an install-license tool that activates this license. From within the distribution's site-packages folder, after navigating to the "cdata/installlic_cosmosdb" folder, simply use a command like the below to activate the license. Omitting the <key> argument activates a trial license:

• Windows:
  

  ./install-license.exe <key>

• Linux / Mac:
  

  ./install-license.sh <key>

Sometimes, file access issues may cause pip to install the connector in a fallback file path that is not the python distribution's main or primary site-packages location. This can make it difficult to find where the connector was installed, and from there, the license activator. In that event, this python script below will print out the full file path of the connector's native file. This file will be stored in the mentioned cdata folder, from which the installlic_cosmosdb folder is trivial to find:

import os
import cdata.cosmosdb
path = os.path.abspath(cdata.cosmosdb.__file__)
print(path)

Uninstallation

If the connector needs to be uninstalled for any reason, do so by running the pip uninstall command, as in the example below:

pip uninstall cdata-cosmosdb-connector

The objects available within our connector are accessible from the "cdata.cosmosdb" module. To use the module's objects directly:

1. Import the module as follows:
   

   import cdata.cosmosdb as mod

2. To establish a connection string, call the connect() method from the connector object using an appropriate connection string, such as:
   

   mod.connect("AccountEndpoint=myAccountEndpoint;AccountKey=myAccountKey;")

Connecting to Azure Cosmos DB

Account Key

Log in to the Azure Portal, select Azure Cosmos DB, and select your account.

Set the following to authenticate:

• AccountEndpoint: The Cosmos DB account URL. Set this to the URI value found in the Settings > Keys blade of the Cosmos DB account.
• AccountKey: A master key token or a resource token for connecting to Azure Cosmos DB. Set this to the PRIMARY KEY value found in the Settings > Keys blade of the Cosmos DB account.
• TokenType: (optional). Set this to "master" (the default value) if you are using a Master Token, which is a full permissions token generated during account creation. Otherwise, set this property to "resource" if you are using a Resource Token, which is a custom permissions token generated when a database user is set up.

Entra ID (Azure AD)

Note: Microsoft has rebranded Azure AD as Entra ID. In topics that require the user to interact with the Entra ID Admin site, we use the same names Microsoft does. However, there are still CData connection properties whose names or values reference "Azure AD".

Microsoft Entra ID is a multi-tenant, cloud-based identity and access management platform. It supports OAuth-based authentication flows that enable the driver to access Azure Cosmos DB endpoints securely.

Authentication to Entra ID via a web application always requires that you first create and register a custom OAuth application. This enables your application to define its own redirect URI, manage credential scope, and comply with organization-specific security policies.

For full instructions on how to create and register a custom OAuth application, see Creating an Entra ID (Azure AD) Application.

After setting AuthScheme to AzureAD, the steps to authenticate vary, depending on the environment. For details on how to connect from desktop applications, web-based workflows, or headless systems, see the following sections.

Desktop Applications

You can authenticate from a desktop application using either the driver's embedded OAuth application or a custom OAuth application registered in Microsoft Entra ID.

Option 1: Use the Embedded OAuth Application

This is a pre-registered application, included with the driver. It simplifies setup and eliminates the need to register your own credentials and is ideal for development environments, single-user tools, or any setup where quick and easy authentication is preferred.

Set the following connection properties:

• AuthScheme: AzureAD
• InitiateOAuth:
  • GETANDREFRESH – Use for the initial login. Launches the login page and saves tokens.
  • REFRESH – Use this setting when you have already obtained valid access and refresh tokens. Reuses stored tokens without prompting the user again.

When you connect, the driver opens the Microsoft Entra sign-in page in your default browser. After signing in and granting access, the driver retrieves the access and refresh tokens and saves them to the path specified by OAuthSettingsLocation.

Option 2: Use a Custom OAuth Application

If your organization requires more control, such as managing security policies, redirect URIs, or application branding, you can instead register a custom OAuth application in Microsoft Entra ID and provide its values during connection.

During registration, record the following values:

• OAuthClientId: The client Id that was generated when you registered your custom OAuth application.
• OAuthClientSecret: The client secret that was that was generated when you registered your custom OAuth application.
• CallbackURL: A redirect URI you defined during application registration.

For full instructions on how to register a custom OAuth application and configure redirect URIs, see Creating an Entra ID (Azure AD) Application.

Set the following connection properties:

• AuthScheme: AzureAD
• InitiateOAuth:
  • GETANDREFRESH – Use for the initial login. Launches the login page and saves tokens.
  • REFRESH – Use this setting when you have already obtained valid access and refresh tokens. Reuses stored tokens without prompting the user again.
• OAuthClientId: The client Id that was generated when you registered your custom OAuth application.
• OAuthClientSecret: The client secret that was generated when you registered your custom OAuth application.
• CallbackURL: A redirect URI you defined during application registration.

After authentication, tokens are saved to OAuthSettingsLocation. These values persist across sessions and are used to automatically refresh the access token when it expires, so you don't need to log in again on future connections.

Web Applications

To authenticate from a web application, you must register a custom OAuth application in Microsoft Entra ID (formerly Azure Active Directory). Embedded OAuth apps are not supported in this context because web-based flows require a registered redirect URI and centralized credential management.

This approach is designed for hosted, multi-user environments where access must be delegated through a secure, standards-compliant OAuth workflow. It gives your organization control over the OAuth client, redirect URI, branding, and permissions scope.

Before you begin: Register a custom OAuth application in the Azure portal. During registration, collect the following values:

• OAuthClientId: The client Id that was generated when you registered your custom OAuth application.
• OAuthClientSecret: The client secret that was generated when you registered your custom OAuth application.
• CallbackURL: A redirect URI you defined during application registration.

For full instructions on how to register a custom OAuth application and configure redirect URIs, see Creating an Entra ID (Azure AD) Application.

To authenticate using AzureAD in a web application, configure the following connection properties:

• AuthScheme: AzureAD
• InitiateOAuth: OFF – Disables automatic login prompts.
• OAuthClientId: The client Id that was generated when you registered your custom OAuth application.
• OAuthClientSecret: The client secret that was generated when you registered your custom OAuth application.
• CallbackURL: A redirect URI you defined during application registration.

Because web applications typically manage OAuth flows manually on the server-side, InitiateOAuth must be set to OFF. This allows you to explicitly control when and how tokens are retrieved and exchanged using stored procedures.

After configuring these properties, follow the steps below to obtain and exchange OAuth tokens:

1. Call the GetOAuthAuthorizationURL stored procedure:
   • CallbackURL: Set to your registered redirect URI
2. Open the returned URL in a browser. Sign in with a Microsoft Entra ID account and grant access.
3. After signing in, you are redirected to your CallbackURL with a code parameter in the query string.
4. Extract the code and pass it to the GetOAuthAccessToken stored procedure:
   • AuthMode: WEB
   • Verifier: The authorization code from the CallbackURL
5. The procedure returns:
   • OAuthAccessToken: Used for authentication.
   • OAuthRefreshToken: Used to refresh the access token.
   • ExpiresIn: The lifetime of the access token in seconds.

To enable automatic token refresh, configure the following connection properties:

• InitiateOAuth: REFRESH
• OAuthRefreshToken: Your previously obtained refresh token

When InitiateOAuth is set to REFRESH, the driver uses the provided refresh token to request a new access token automatically.

After a successful connection, the driver saves the updated access and refresh tokens to the file specified by OAuthSettingsLocation.

You only need to repeat the full OAuth authorization flow if the refresh token expires, is revoked, or becomes invalid.

For more background on OAuth flows in Microsoft Entra ID, see Microsoft Entra Authentication Overview.

Headless Machines

Headless environments like CI/CD pipelines, background services, or server-based integrations do not have an interactive browser. To authenticate using AzureAD, you must complete the OAuth flow on a separate device with a browser and transfer the authentication result to the headless system.

Setup options:

• Obtain and exchange a verifier code
  • Use another device to sign in and retrieve a verifier code, which the headless system uses to request tokens.
• Transfer an OAuth settings file
  • Authenticate on another device, then copy the stored token file to the headless environment.

Using a Verifier Code

1. On a device with a browser:
   • If using a custom OAuth app, set the following properties:
     • InitiateOAuth: OFF
     • OAuthClientId: The client Id that was generated when you registered your custom OAuth application.
     • OAuthClientSecret: The client secret that was generated when you registered your custom OAuth application.
   • Call the GetOAuthAuthorizationURL stored procedure to generate a sign-in URL.
   • Open the returned URL in a browser. Sign in and grant permissions to the driver. You are redirected to the callback URL, which contains the verifier code.
   • After signing in, save the value of the code parameter from the redirect URL. You will use this later to set the OAuthVerifier connection property.
2. On the headless machine:
   • Set the following properties:
     • AuthScheme: AzureAD
     • InitiateOAuth: REFRESH
     • OAuthVerifier: The verifier code you saved.
     • OAuthSettingsLocation: The path of the file that holds the OAuth token values.
     • For custom applications:
       • OAuthClientId: The client Id that was generated when you registered your custom OAuth application.
       • OAuthClientSecret: The client secret that was generated when you registered your custom OAuth application.
   • After tokens are saved, reuse them by setting:
     • InitiateOAuth: REFRESH
     • OAuthSettingsLocation: Make sure this location grants read and write permissions to the driver to enable the automatic refreshing of the access token.
     • For custom applications:
       • OAuthClientId: The client Id that was generated when you registered your custom OAuth application.
       • OAuthClientSecret: The client secret that was generated when you registered your custom OAuth application.

Transferring OAuth Settings

1. On a device with a browser:
   • Connect using the instructions in the Desktop Applications section.
   • After connecting, tokens are saved to the file path in OAuthSettingsLocation. The default filename is OAuthSettings.txt.

2. On the headless machine:
   • Copy the OAuth settings file to the machine.
   • Set the following properties:
     • AuthScheme: AzureAD
     • InitiateOAuth: REFRESH
     • OAuthSettingsLocation: Make sure this location grants read and write permissions to the driver to enable the automatic refreshing of the access token.
     • For custom applications:
       • OAuthClientId: The client Id that was generated when you registered your custom OAuth application.
       • OAuthClientSecret: The client secret that was generated when you registered your custom OAuth application.

After setup, the driver uses the stored tokens to refresh the access token automatically, no browser or manual login is required.

Azure Service Principal

Note: Microsoft has rebranded Azure AD as Entra ID. In topics that require the user to interact with the Entra ID Admin site, we use the same names Microsoft does. However, there are still CData connection properties whose names or values reference "Azure AD".

Azure Service Principal is role-based application-based authentication. This means that authentication is done per application, rather than per user. All tasks taken on by the application are executed without a default user context, but based on the assigned roles. The application access to the resources is controlled through the assigned roles' permissions.

For information about how to set up Azure Service Principal authentication, see Creating a Service Principal App in Entra ID (Azure AD).

Java Native Interface (JNI) is a standard programming interface for writing Java native methods and embedding the Java virtual machine into native applications.

The connector leverages the JNI for improved performance on Mac and Linux.

Configure the Config INI File

The Linux and Mac editions of the Azure Cosmos DB python connector are configured with an ini file. This file is used to set several parameters, including JNI behavior. This file is to be located in:

{path_to_distribution_site-packages}/cdata/config.ini

Ensure that any configuration properties you set in the ini file fall under the following section name (adjust the 311 number if you are using an different python version from 3.11):

• For Linux:
  

  [cosmosdb.cpython-311-x86_64-linux-gnu.so]

• For Mac:
  

  [cosmosdb.cpython-311-darwin.so]

Configure the JNI connector's behavior by editing the properties in the connector's config.ini file. The connector can be configured as follows:

• LOGFILE: Set this the same way as the CDATA_LOGFILE envrionment variable below.
• JAVA_HOME: Configure the path to the JVM library location used to launch the JVM.
• CLASS_PATH: Use a colon-separated list to configure the paths to the third-party jar libraries.

Configure Environment Variables

Additionally, set the following environment variables:

• CDATA_JAVA_HOME: Configure the path to the JVM library location used to launch the JVM.
• CDATA_JVM_OPTIONS: Place JVM options here.
• CDATA_LOGFILE: Set this in the following scheme: <SCHEME>://<TAG>[|<LEVEL>]
  • SCHEME: The options are STDOUT, FILE.
    • STDOUT: Both the native wrapper and odbc core log into stdout. The Logfile and Verbosity properties can override the behavior of ODBC core.
    • FILE: The native wrapper logs into <FILENAME> while the odbc core logs into <FILENAME>.driver.log. The Logfile and Verbosity properties can override the behavior of ODBC core.
  • TAG
    • For STDOUT, set this to 1. For FILE, set this to the filename.
  • LEVEL
    • Set to one of: FATAL | ERROR | WARNING | INFO | DEBUG

The following are some examples of this syntax:

• STDOUT://1|DEBUG
• FILE:///tmp/my_py.log|DEBUG

Custom Logger

The Python connector supports a custom logging mechanism for redirecting log output to any destination, such as a cloud storage service or logging framework. Use setCustomLoggerFactory() to register a factory function that creates a logger instance for each connection.

The factory function receives the context string from the Logfile connection property (the portion after CUSTOM://) and must return an object with a writeLog(verbosity, message) method.

To enable custom logging:

1. Call setCustomLoggerFactory() with your factory function before opening connections.
2. Set Logfile to CUSTOM:// followed by a context string to identify the connection.
3. Set Verbosity to the desired log level.

The following example demonstrates a custom logger factory that creates a separate logger instance per connection:

import cdata.cosmosdb as mod
import time

class MyLogger:
    def __init__(self, loggerId):
        self.loggerId = loggerId
    def writeLog(self, verbosity, message):
        print("[MyLogger " + self.loggerId + "] " + message)

def createLogger(context):
    return MyLogger(context[len("MyLoggerId="):])

mod.setCustomLoggerFactory(createLogger)

conn1 = mod.connect("...;Logfile=CUSTOM://MyLoggerId=1;Verbosity=2;")
# do something with conn1
time.sleep(1)  # Wait for logs to flush from conn1

conn2 = mod.connect("...;Logfile=CUSTOM://MyLoggerId=2;Verbosity=2;")
# do something with conn2
time.sleep(1)  # Wait for logs to flush from conn2

Creating an Entra ID (Azure AD) Application

Note: Microsoft has rebranded Azure AD as Entra ID. In topics that require the user to interact with the Entra ID Admin site, we use the same names Microsoft does. However, there are still CData connection properties whose names or values reference "Azure AD".

Azure Cosmos DB supports OAuth-based authentication using Microsoft Entra ID. If you will connect via a web application and want to authenticate via Entra ID, you must first register a custom OAuth application in the Entra Admin Center, as described below.

Registering the Application

To register an OAuth application in Microsoft Entra ID, follow these steps:

1. Go to https://portal.azure.com.
2. In the left-hand navigation pane, select Microsoft Entra ID > App registrations.
3. Click New registration.
4. Enter a name for the application.
5. Specify the types of accounts this application should support:
   • For private-use applications, select Accounts in this organization directory only.
   • For distributed applications, select one of the multi-tenant options.

   Note: If you select Accounts in this organizational directory only, when you connect with CData Python Connector for Azure Cosmos DB, you must set AzureTenant to the tenant's ID (either GUID or verified domain). Otherwise, authentication will fail.

6. Set Select a platform to Web, and set the redirect URI to http://localhost:33333 (default), or use another URI appropriate for your deployment. When using a custom redirect URI set a CallbackURL connection property; in those cases, set it to match this URI exactly.
7. Click Register. The application management screen opens. Record these values for later use:
   • Application (client) ID is used for OAuthClientId
   • Directory (tenant) ID is used for AzureTenant
8. Go to Certificates & Secrets. Click New Client Secret, set the desired expiration, and save the generated value. This value will only be shown once — record it to use with OAuthClientSecret.

9. To confirm, click Add permissions.

Creating a Service Principal App in Entra ID (Azure AD)

Note: Microsoft has rebranded Azure AD as Entra ID. In topics that require the user to interact with the Entra ID Admin site, we use the same names Microsoft does. However, there are still CData connection properties whose names or values reference "Azure AD".

Azure Cosmos DB supports Service Principal-based authentication, which is role-based. This means that the Service Principal's permissions are determined by the roles assigned to it. The roles specify what resources the Service Principal can access and which operations it can perform.

If you want to use a Service Principal to authenticate to Azure Cosmos DB, you must create a custom application in Microsoft Entra ID.

To enable Service Principal authentication:

• Confirm that you have permission to register applications and assign roles in your tenant.
• Register a new application and configure credentials and permissions in the Entra Admin Center.

Registering the Application

1. Go to https://portal.azure.com.
2. In the left-hand navigation pane, select Microsoft Entra ID > App registrations.
3. Click New registration.
4. Enter a name for the application.
5. Select the desired tenant setup. Since this custom application is for Service Principal use, choose Any Microsoft Entra ID tenant – Multitenant.

6. Click Register. The application management screen opens. Note the value in Application (client) ID as the OAuthClientId and the Directory (tenant) ID as the AzureTenant

7. Navigate to Certificates & Secrets and define the application authentication type. Two types of authentication are available: certificate (recommended) or client secret
   • For certificate authentication: In Certificates & Secrets, select Upload certificate, then upload the certificate from your local machine. For more information on creating a self-signed certificate, see Create a self-signed certificate
   • For creating a new client secret: In Certificates & Secrets, select New Client Secret for the application and specify its duration. After the client secret is saved, Azure Cosmos DB displays the key value. This value is displayed only once, so be sure to record it for future use. Use this value for the OAuthClientSecret

8. Navigate to Authentication and select the Access tokens option.
9. Save your changes.

Consent for Client Credentials

Note: Since the embedded OAuth credentials authenticate on a per-user basis, you cannot use them in a client authentication flow. You must always create a custom OAuth application to use client credentials.

1. Log in to https://portal.azure.com
2. Create a custom OAuth application, as described above.
3. Navigate to App Registrations.
4. Find the application you just created, and open API Permissions.
5. Select the Microsoft Graph permissions. There are two distinct sets of permissions: Delegated and Application.
6. For use with Service Principal, specify Application permissions.
7. Select any additional permissions you require for your integration.

Fine Tuning Data Access

You can use the following properties to gain greater control over Azure Cosmos DB API features and the strategies the connector uses to surface them:

• RowScanDepth: This property determines the number of rows that will be scanned to detect column data types when generating table metadata.
• TypeDetectionScheme: This property allows more control over the strategy implemented by the RowScanDepth property.

• GenerateSchemaFiles: This property enables you to persist table metadata in static schema files that are easy to customize, to persist your changes to column data types, for example.
  You can set this property to "OnStart" to generate schema files for all tables in your database at connection. Or, you can generate schemas as you execute SELECT queries to tables.
  The resulting schemas are based on the connection properties you use to configure Automatic Schema Discovery
  To use the resulting schema files, set the Location property to the folder containing the schemas.

Just as described in the SQL Compliance the connector supports batch CUD (Create, Update, Delete) operations. Batch processing is achieved by issuing multiple requests simultaneously. Even though this method greatly improves the performance for write operations, the cost of these operations is relatively high, thus the Request Units (RU) budget per second for a certain container or database may be exceeded. Depending on your Azure Cosmos DB Service Quotas, exceeding the RU budgets may incur in extra costs, or it may even temporary throttle or interrupt the Azure Cosmos DB usage for other workloads.

In order to avoid exceeding the RU budget per second, the connector dynamically adjusts the number of concurrent requests per second depending on the set WriteThroughputBudget and the constantly adjusted average RU cost per statement. The user can utilize the WriteThroughputBudget connection property to define the RU budged per second, that batch write operations should not exceed. Another important factor in batch write operations is the MaxThreads connection property, which specifies the maximum number of concurrent requests. If using a low MaxThreads value, the connector might not be able to efficiently use the available budget.

Since the requests throttling logic is applied client-side, in a few cases the RU/s budged may be exceeded by a relatively small amount. These cases include inserting, updating and deleting records with highly variable column count and input value length per column.

Note: By default, the WriteThroughputBudget property is set 1000 RU/s and the MaxThreads property is set to 200 threads.

General Changes

Azure Cosmos DB is a schemaless, document database that provides high performance, availability, and scalability. These features are not necessarily incompatible with a standards-compliant query language like SQL-92. In this section we will show various schemes that the connector offers to bridge the gap with relational SQL and a document database.

Working with Azure Cosmos DB Objects as Tables

The connector models the schemaless Azure Cosmos DB objects into relational tables and translates SQL queries into Azure Cosmos DB queries to get the requested data. See Query Mapping (Sql API) for more details on how various Azure Cosmos DB operations are represented as SQL.

Discovering Schemas Automatically

The Automatic Schema Discovery scheme automatically finds the data types in a Azure Cosmos DB object by scanning a configured number of rows of the object. You can use RowScanDepth, FlattenArrays, and FlattenObjects to control the relational representation of the collections in Azure Cosmos DB. You can also write Free-Form Queries not tied to the schema.

Customizing Schemas

Optionally, you can use Custom Schema Definitions to project your chosen relational structure on top of a Azure Cosmos DB object. This allows you to define your chosen names of columns, their data types, and the location of their values in the collection.

Set GenerateSchemaFiles to save the detected schemas as simple configuration files that are easy to extend. You can persist schemas for all collections in the database or for the results of SELECT queries.

Limitations of the RawValue TypeDetectionScheme

If the TypeDetectionScheme is set to RawValue, the connector will push each document as single aggregate value on a column named JsonData, along with its resource identifier on the separate Primary Key column. The JSON documents are not processed, and as a result, the below functionalities are NOT supported with this configuration.

• Automatic Schema Discovery
• Free-Form Queries
• Vertical Flattening
• SQL API Built-In Functions
• SQL API GROUP BY
• SQL API JOIN IN
• Almost all server side supported filters apart from WHERE clause conditions built with the resource identifier.

The connector automatically infers a relational schema by inspecting a series of Azure Cosmos DB documents in a collection. You can use the RowScanDepth property to define the number of documents the connector will scan to do so. The columns identified during the discovery process depend on the FlattenArrays and FlattenObjects properties.

Flattening Objects

If FlattenObjects is set, all nested objects will be flattened into a series of columns. For example, consider the following document:

{
  id: 12,
  name: "Lohia Manufacturers Inc.",
  address: {street: "Main Street", city: "Chapel Hill", state: "NC"},
  offices: ["Chapel Hill", "London", "New York"],
  annual_revenue: 35,600,000
}

If FlattenObjects is not set, then the address.street, address.city, and address.state columns will not be broken apart. The address column of type string will instead represent the entire object. Its value would be {street: "Main Street", city: "Chapel Hill", state: "NC"}. See JSON Functions for more details on working with JSON aggregates.

You can change the separator character in the column name from a dot by setting SeparatorCharacter.

Flattening Arrays

The FlattenArrays property can be used to flatten array values into columns of their own. This is only recommended for arrays that are expected to be short, for example the coordinates below:

"coord": [ -73.856077, 40.848447 ]

It is best to leave other unbounded arrays as they are and piece out the data for them as needed using JSON Functions.

As discussed in Automatic Schema Discovery, intuited table schemas enable SQL access to unstructured Azure Cosmos DB data. JSON Functions enable you to use standard JSON functions to summarize Azure Cosmos DB data and extract values from any nested structures. Custom Schema Definitions enable you to define static tables and give you more granular control over the relational view of your data; for example, you can write schemas defining parent/child tables or fact/dimension tables. However, you are not limited to these schemes.

After connecting you can query any nested structure without flattening the data. Any relations that you can access with FlattenArrays and FlattenObjects can also be accessed with an ad hoc SQL query.

Let's consider an example document from the following Restaurant data set:

 
{
  "address": {
    "building": "1007",
    "coord": [
      -73.856077,
      40.848447
    ],
    "street": "Morris Park Ave",
    "zipcode": "10462"
  },
  "borough": "Bronx",
  "cuisine": "Bakery",
  "grades": [
    {
      "grade": "A",
      "score": 2,
      "date": {
        "$date": "1393804800000"
      }
    },
    {
      "date": {
        "$date": "1378857600000"
      },
      "grade": "B",
      "score": 6
    },
    {
      "score": 10,
      "date": {
        "$date": "1358985600000"
      },
      "grade": "C"
    }
  ],
  "name": "Morris Park Bake Shop",
  "restaurant_id": "30075445"
} 

SELECT [address.building], [grades.1.grade] FROM restaurants WHERE restaurant_id = '30075445'

It is possible to retrieve an array of documents as if it were a separate table. Take the following JSON structure from the restaurants collection for example:

{
  "_id" : ObjectId("568c37b748ddf53c5ed98932"),
  "address" : {
    "building" : "1007",
    "coord" : [-73.856077, 40.848447],
    "street" : "Morris Park Ave",
    "zipcode" : "10462"
  },
  "borough" : "Bronx",
  "cuisine" : "Bakery",
  "grades" : [{
      "date" : ISODate("2014-03-03T00:00:00Z"),
      "grade" : "A",
      "score" : 2
    }, {
      "date" : ISODate("2013-09-11T00:00:00Z"),
      "grade" : "A",
      "score" : 6
    }, {
      "date" : ISODate("2013-01-24T00:00:00Z"),
      "grade" : "A",
      "score" : 10
    }, {
      "date" : ISODate("2011-11-23T00:00:00Z"),
      "grade" : "A",
      "score" : 9
    }, {
      "date" : ISODate("2011-03-10T00:00:00Z"),
      "grade" : "B",
      "score" : 14
    }],
  "name" : "Morris Park Bake Shop",
  "restaurant_id" : "30075445"
}

SELECT * FROM [restaurants.grades]

You may also want to include information from the base restaurants table. You can do this with a join. Flattened arrays can only be joined with the root document. The connector expects the left part of the join is the array document you want to flatten vertically. Set SupportEnhancedSQL to false to join nested Azure Cosmos DB documents -- this type of query is supported through the Azure Cosmos DB API.

SELECT [restaurants].[restaurant_id], [restaurants.grades].* FROM [restaurants.grades] JOIN [restaurants] WHERE [restaurants].name = 'Morris Park Bake Shop'

The connector can return JSON structures as column values. The connector enables you to use standard SQL functions to work with these JSON structures. The examples in this section use the following array:

[
     { "grade": "A", "score": 2 },
     { "grade": "A", "score": 6 },
     { "grade": "A", "score": 10 },
     { "grade": "A", "score": 9 },
     { "grade": "B", "score": 14 }
]

JSON_EXTRACT

SELECT Name, JSON_EXTRACT(grades,'[0].grade') AS Grade, JSON_EXTRACT(grades,'[0].score') AS Score FROM Students;

JSON_COUNT

SELECT Name, JSON_COUNT(grades,'[x]') AS NumberOfGrades FROM Students;

JSON_SUM

SELECT Name, JSON_SUM(score,'[x].score') AS TotalScore FROM Students;

JSON_MIN

SELECT Name, JSON_MIN(score,'[x].score') AS LowestScore FROM Students;

JSON_MAX

SELECT Name, JSON_MAX(score,'[x].score') AS HighestScore FROM Students;

DOCUMENT

The DOCUMENT function can be used to retrieve the entire document as a JSON string. See the following query and its result as an example:

SELECT DOCUMENT(*) FROM Customers;

{ "id": 12, "name": "Lohia Manufacturers Inc.", "address": { "street": "Main Street", "city": "Chapel Hill", "state": "NC"}, "offices": [ "Chapel Hill", "London", "New York" ], "annual_revenue": 35,600,000 }

Cosmos DB also supports a number of built-in functions for common operations, that can be used inside queries. Here are some example of how can be used as part of select columns or the WHERE clause:

Use Built-in functions as part of SELECT columns

SELECT IS_NUMBER(user_id) AS ISN_ATTR, IS_NUMBER(id) AS ISN_ID FROM [users]
SELECT POWER(user_id, 2) AS POWERSSS, LENGTH(id) AS LENGTH_ID, PI() AS JustThePI FROM [users]

Use Built-in functions as part of WHERE clause

SELECT * FROM [users] WHERE STARTSWITH(middle_name, 'G')
SELECT * FROM [users] WHERE REPLACE(middle_name, 'Chr', '___') = '___istopher'

Mathematical functions

The mathematical functions each perform a calculation, based on input values that are provided as arguments, and return a numeric value. Here's a table of supported built-in mathematical functions.

Type checking functions

The type checking functions allow you to check the type of an expression within SQL queries. Type checking functions can be used to determine the type of properties within documents dynamically when it is variable or unknown. Here's a table of supported built-in type checking functions.

String functions

The following scalar functions perform an operation on a string input value and return a string, numeric or Boolean value. Here's a table of built-in string functions:

Array functions

The following scalar functions perform an operation on an array input value and return numeric, Boolean or array value. Here's a table of built-in array functions:

Nested functions

You can also perform nested built-in functions, which are processed server side as well:

i.e. SELECT TOP 10 CONCAT(SUBSTRING(UPPER(cuisine), 0, 3), '-cuisine') FROM [restaurants]

The GROUP BY clause divides the query's results according to the values of one or more specified properties. This operation is partially done server-side because of some API limitations. We still need to operate a client-side grouping.

GROUP BY Examples

SELECT COUNT(*) AS CNT, gender FROM [users] GROUP BY gender
SELECT COUNT(*) AS CNT, gender, doc_type FROM [users] GROUP BY gender, doc_type

Cosmos DB's SQL API supports a special type of join operation called JOIN IN, which is specifically designed for working with nested arrays within documents. Unlike traditional SQL joins that combine data from separate tables, JOIN IN allows you to "flatten" and query nested array elements within a single document.

Document Structure Example

        {
            "id": "3",
            "name": "DEV Park Bake Shop",
            "cuisine": "Bakery",
            "grades": [
                {
                    "date": 1393804800000,
                    "grade": "D",
                    "score": 2
                },
                {
                    "date": 1378857600000,
                    "grade": "A",
                    "score": 6
                }
            ]
        }
    

SQL Query Syntax

        SELECT c.Id, g.grade, g.score, g.date
        FROM restaurants c
        JOIN g IN c.grades
        WHERE c.[name] = 'DEV Park Bake Shop'
    

CosmosDB Translation

        SELECT c["Id"], g["grade"], g["score"], g["date"]
        FROM C AS c
        JOIN g IN c.grades
        WHERE c["name"] = "DEV Park Bake Shop"
    

The connector maps SQL queries into the corresponding Azure Cosmos DB SQL API queries. A detailed description of all the transformations is out of scope, but we will describe some of the common elements that are used. The connector takes advantage of SQL API features such as the aggregation framework to compute the desired results.

SELECT Queries

SELECT id, name FROM Users

SELECT C.id, C.name FROM C

SELECT * FROM Users WHERE name = 'A'

SELECT * FROM C WHERE C.name = 'A'

SELECT * FROM Users WHERE name = 'A' OR email = 'zoe55@gmail.com'

SELECT * FROM C WHERE C.name = 'A' OR C.email = 'zoe55@gmail.com'

SELECT id, grantamt FROM WorldBank WHERE grantamt IN (4500000, 85400000) OR grantamt = 16200000

SELECT C.id, C.grantamt FROM C WHERE C.grantamt IN (4500000, 85400000) OR C.grantamt = 16200000

SELECT * FROM WorldBank WHERE CountryCode = 'A' ORDER BY TotalCommAmt ASC

SELECT * FROM C WHERE C.countrycode = 'AL' ORDER BY C.totalcommamt ASC

SELECT * FROM WorldBank WHERE CountryCode = 'A' ORDER BY TotalCommAmt DESC

SELECT * FROM C WHERE C.countrycode = 'AL' ORDER BY C.totalcommamt DESC

Aggregate Queries

SELECT COUNT(grantamt) AS COUNT_GRAMT FROM WorldBank

SELECT COUNT(C.grantamt) AS COUNT_GRAMT FROM C

SELECT SUM(grantamt) AS SUM_GRAMT FROM WorldBank

SELECT SUM(C.grantamt) AS SUM_GRAMT FROM C

Built-In functions

SELECT IS_NUMBER(grantamt) AS ISN_ATTR, IS_NUMBER(id) AS ISN_ID FROM WorldBank

SELECT IS_NUMBER(C.grantamt) AS ISN_ATTR, IS_NUMBER(C.id) AS ISN_ID FROM C

SELECT POWER(totalamt, 2) AS POWERS_A, LENGTH(id) AS LENGTH_ID, PI() AS ThePI FROM WorldBank

SELECT POWER(C.totalamt, 2) AS POWERS_A, LENGTH(C.id) AS LENGTH_ID, PI() AS ThePI FROM C

You can extend the table schemas created with Automatic Schema Discovery by saving them into schema files. The schema files have a simple format that makes the schemas to edit.

Generating Schema Files

Set GenerateSchemaFiles to "OnStart" to persist schemas for all tables when you connect. You can also generate table schemas as needed: Set GenerateSchemaFiles to "OnUse" and execute a SELECT query to the table.

For example, consider a schema for the restaurants data set. This is a sample data set provided by Azure Cosmos DB.

Below is an example document from the collection:

{
  "address":{
    "building":"461",
      "coord":[
        -74.138492,
        40.631136
      ],
      "street":"Port Richmond Ave",
      "zipcode":"10302"
   },
   "borough":"Staten Island",
   "cuisine":"Other",
   "name":"Indian Oven",
   "restaurant_id":"50018994"
}

Customizing a Schema

When GenerateSchemaFiles is set, the connector saves schemas into the folder specified by the Location property. You can then change column behavior in the resulting schema.

The following schema uses the other:bsonpath property to define where in the collection to retrieve the data for a particular column. Using this model you can flatten arbitrary levels of hierarchy.

Below are the corresponding column definitions for the restaurants data set. In Custom Schema Example, you will find the complete schema.

<rsb:script xmlns:rsb="http://www.rssbus.com/ns/rsbscript/2">  

  <rsb:info title="StaticRestaurants" description="Custom Schema for the restaurants data set.">  
    <!-- Column definitions -->
    <attr   name="_rid"               xs:type="string"   key="true"   other:collrid="hWdRAKRi3Pg=" other:dbrid="hWdRAA==" other:partitionpath="/name" />
	<attr   name="borough"            xs:type="string"   />
    <attr   name="cuisine"            xs:type="string"   />
    <attr   name="address.building"   xs:type="string"   />
    <attr   name="address.street"     xs:type="string"   />
    <attr   name="address.coord.0"    xs:type="double"   />
    <attr   name="address.coord.1"    xs:type="double"   />
    <input name="rows@next" desc="Internal attribute used for paging through data."  />
  </rsb:info>  

  <rsb:set attr="collection" value="restaurants"/>

</rsb:script>

This section contains a complete schema. The info section enables a relational view of a Azure Cosmos DB object. For more details, see Custom Schema Definitions. The table below allows the SELECT, INSERT, UPDATE, and DELETE commands as implemented in the GET, POST, MERGE, and DELETE sections of the schema below.

Copy the rows@next input as-is into your schema. The operations, such as cosmosdbadoSysData, are internal implementations and can also be copied as is.

Set the Location property to the file directory that will contain the schema file.

When, creating custom schemas, the attr for _rid, shown below, is required.

Also required are three properties for the _rid column definition:

• other:dbrid is found in the _self property of an item in the collection, after "dbs/".
• other:collrid is found in the _self property of an item in the collection, after "/colls/".
• other:partitionpath refers to the name of the partition specified when the collection was created.

<rsb:script xmlns:rsb="http://www.rssbus.com/ns/rsbscript/2">  

  <rsb:info title="StaticRestaurants" description="Custom Schema for the restaurants data set.">  
    <!-- Column definitions -->
	<attr   name="_rid"               xs:type="string"   key="true"   other:collrid="hWdRAKRi3Pg=" other:dbrid="hWdRAA==" other:partitionpath="/name" />
    <attr   name="borough"            xs:type="string"   />
    <attr   name="cuisine"            xs:type="string"   />
    <attr   name="address.building"   xs:type="string"   />
    <attr   name="address.street"     xs:type="string"   />
    <attr   name="address.coord.0"    xs:type="double"   />
    <attr   name="address.coord.1"    xs:type="double"   />
    <input name="rows@next" desc="Internal attribute used for paging through data."  />
  </rsb:info>  

  <rsb:script method="GET">
    <rsb:call op="cosmosdbadoSysData">
      <rsb:push />
    </rsb:call>
  </rsb:script>

  <rsb:script method="POST">
    <rsb:call op="cosmosdbadoSysData">
      <rsb:push />
    </rsb:call>
  </rsb:script>

  <rsb:script method="MERGE">
    <rsb:call op="cosmosdbadoSysData">
      <rsb:push />
    </rsb:call>
  </rsb:script>

  <rsb:script method="DELETE">
    <rsb:call op="cosmosdbadoSysData">
      <rsb:push />
    </rsb:call>
  </rsb:script>

</rsb:script>

You can query the system tables described in this section to access schema information, information on data source functionality, and batch operation statistics.

Schema Tables

The following tables return database metadata for Azure Cosmos DB:

• sys_catalogs: Lists the available databases.
• sys_schemas: Lists the available schemas.
• sys_tables: Lists the available tables and views.
• sys_tablecolumns: Describes the columns of the available tables and views.
• sys_procedures: Describes the available stored procedures.
• sys_procedureparameters: Describes stored procedure parameters.
• sys_keycolumns: Describes the primary and foreign keys.
• sys_indexes: Describes the available indexes.

Data Source Tables

The following tables return information about how to connect to and query the data source:

• sys_connection_props: Returns information on the available connection properties.
• sys_sqlinfo: Describes the SELECT queries that the connector can offload to the data source.

Query Information Tables

The following table returns query statistics for data modification queries, including batch operations:

• sys_identity: Returns information about batch operations or single updates.

Lists the available databases.

The following query retrieves all databases determined by the connection string:

SELECT * FROM sys_catalogs

Columns

Lists the available schemas.

The following query retrieves all available schemas:

          SELECT * FROM sys_schemas
          

Columns

Lists the available tables.

The following query retrieves the available tables and views:

          SELECT * FROM sys_tables
          

Columns

Describes the columns of the available tables and views.

The following query returns the columns and data types for the [CData].[Entities].Customers table:

SELECT ColumnName, DataTypeName FROM sys_tablecolumns WHERE TableName = 'Customers' AND CatalogName = 'CData' AND SchemaName = 'Entities'

Columns

Lists the available stored procedures.

The following query retrieves the available stored procedures:

          SELECT * FROM sys_procedures
          

Columns

Describes stored procedure parameters.

The following query returns information about all of the input parameters for the EVAL stored procedure:

SELECT * FROM sys_procedureparameters WHERE ProcedureName = 'EVAL' AND Direction = 1 OR Direction = 2

To include result set columns in addition to the parameters, set the IncludeResultColumns pseudo column to True:

SELECT * FROM sys_procedureparameters WHERE ProcedureName = 'EVAL' AND IncludeResultColumns='True'

Columns

Pseudo-Columns

Describes the primary and foreign keys.

The following query retrieves the primary key for the [CData].[Entities].Customers table:

         SELECT * FROM sys_keycolumns WHERE IsKey='True' AND TableName='Customers' AND CatalogName='CData' AND SchemaName='Entities'
          

Columns

Describes the foreign keys.

The following query retrieves all foreign keys which refer to other tables:

         SELECT * FROM sys_foreignkeys WHERE ForeignKeyType = 'FOREIGNKEY_TYPE_IMPORT'
          

Columns

Describes the primary keys.

The following query retrieves the primary keys from all tables and views:

         SELECT * FROM sys_primarykeys
          

Columns

Describes the available indexes. By filtering on indexes, you can write more selective queries with faster query response times.

The following query retrieves all indexes that are not primary keys:

          SELECT * FROM sys_indexes WHERE IsPrimary='false'
          

Columns

Returns information on the available connection properties and those set in the connection string.

The following query retrieves all connection properties that have been set in the connection string or set through a default value:

SELECT * FROM sys_connection_props WHERE Value <> ''

Columns

Describes the SELECT query processing that the connector can offload to the data source.

See SQL Compliance for SQL syntax details.

Discovering the Data Source's SELECT Capabilities

Below is an example data set of SQL capabilities. Some aspects of SELECT functionality are returned in a comma-separated list if supported; otherwise, the column contains NO.

The following query retrieves the operators that can be used in the WHERE clause:

SELECT * FROM sys_sqlinfo WHERE Name = 'SUPPORTED_OPERATORS'

Columns

Returns information about attempted modifications.

The following query retrieves the Ids of the modified rows in a batch operation:

         SELECT * FROM sys_identity
          

Columns

Describes the available system information.

The following query retrieves all columns:

SELECT * FROM sys_information

Columns

Stored procedures are function-like interfaces that extend the functionality of the connector beyond simple SELECT/INSERT/UPDATE/DELETE operations with Azure Cosmos DB.

Stored procedures accept a list of parameters, perform their intended function, and then return any relevant response data from Azure Cosmos DB, along with an indication of whether the procedure succeeded or failed.

CData Python Connector for Azure Cosmos DB Stored Procedures

Insert entire JSON string to CosmosDB.

Input

Result Set Columns

Creates a schema file for the collection.

CreateSchema

Creates a local schema file (.rsd) from an existing table or view in the data model.

The schema file is created in the directory set in the Location connection property when this procedure is executed. You can edit the file to include or exclude columns, rename columns, or adjust column datatypes.

The connector checks the Location to determine if the names of any .rsd files match a table or view in the data model. If there is a duplicate, the schema file will take precedence over the default instance of this table in the data model. If a schema file is present in Location that does not match an existing table or view, a new table or view entry is added to the data model of the connector.

Input

Result Set Columns

Gets the OAuth access token from CosmosDB.

Input

Result Set Columns

Gets the CosmosDB authorization URL. Access the URL returned in the output in a Web browser. This requests the access token that can be used as part of the connection string to CosmosDB.

Input

Result Set Columns

Refreshes the OAuth access token used for authentication with CosmosDB.

Input

Result Set Columns

This section provides a walk-through for writing Azure Cosmos DB data access code in Python script.

For more information on the available data source entities and how to query them with SQL, see NoSQL Database. For the SQL syntax, see SQL Compliance.

Connecting from Code

For information on how to connect with the cosmosdb.connector module and its related classes, see Connecting.

Executing SQL

Executing Stored Procedures

Batch Processing

Connecting with the cdata.cosmosdb Module:

import cdata.cosmosdb as mod
conn = mod.connect("AccountEndpoint=myAccountEndpoint;AccountKey=myAccountKey;")

Once the connection is created, you can use it to execute subsequent SQL queries.

After connecting as described in Connecting, you can use the open connection to execute SQL statements.

Executing Queries

For example:

cur = conn.execute("SELECT City, CompanyName FROM [CData].[Entities].Customers")
rs = cur.fetchall()
for row in rs:
	print(row)

Parameterized Queries

Various Python collections, such as arrays and tuples, can act as additional arguments for the execute() method. This enables you to parameterize the queries executed and help to prevent SQL Injection.

For example:

cmd = "SELECT City, CompanyName FROM [CData].[Entities].Customers WHERE Country = ?"
params = ["US"]
cur = conn.execute(cmd, params)
rs = cur.fetchall()
for row in rs:
	print(row)

The connection is also used to issue INSERT, UPDATE, and DELETE commands to the data source. Parameters can be used with these statements if desired.

Note that the connector does not support transactions. As with normal write operations, all SQL statements executed by this connector affect the data source immediately. Call the connection's commit() method following the execution.

Insert

cmd = "INSERT INTO [CData].[Entities].Customers (City, CompanyName) VALUES (?, ?)"
params = ["Jon Deere", "Caterpillar"]
cur = conn.execute(cmd, params)
print("Records affected: ", cur.rowcount)

Update

cmd = "UPDATE [CData].[Entities].Customers SET CompanyName = ? WHERE _id = ?"
params = ["Caterpillar", "22"]
cur = conn.execute(cmd, params)
print("Records affected: ", cur.rowcount)

Delete

The following example removes an existing record from the table:

cmd = "DELETE FROM [CData].[Entities].Customers WHERE _id = ?"
params = ["22"]
cur = conn.execute(cmd, params)
print("Records affected: ", cur.rowcount)

You can execute stored procedures using either the execute() or callproc() method of the connection.

Calling Stored Procedures Using Execute()

cmd = "EXECUTE EVAL jsFunction = ?"
params = ["function () { return db.restaurants.findOne(); }"]
conn.execute(cmd, params)

Calling Stored Procedures Using Callproc()

cur = conn.cursor()
params = ["function () { return db.restaurants.findOne(); }"]
cur.callproc("EVAL", params)

This Python connector also supports writing to the data source via batch processing, using the cursor object's executemany() method. This requires both a SQL statement string and a data frame of values that act as a series of parameters for executing the SQL statement.

Note that the connector does not support transactions. As with normal write operations, all SQL statements executed by this connector affect the data source immediately. Call the connection's commit() method following the execution.

Insert

cur = conn.cursor()
cmd = "INSERT INTO [CData].[Entities].Customers (City, CompanyName) VALUES (?, ?)"
params = [["Jon Deere", "Caterpillar"], ["Jon Deere", "Caterpillar"]]
cur.executemany(cmd, params)
print("Records affected: ", cur.rowcount)

Update

cur = conn.cursor()
cmd = "UPDATE [CData].[Entities].Customers SET CompanyName = ? WHERE _id = ?"
params = [["Caterpillar", "22"], ["Caterpillar", "22"]]
cur.executemany(cmd, params)
print("Records affected: ", cur.rowcount)

Delete

cur = conn.cursor()
cmd = "DELETE FROM [CData].[Entities].Customers WHERE _id = ?"
params = [["22"], ["22"]]
cur.executemany(cmd, params)
print("Records affected: ", cur.rowcount)

The connector is integrated with other tools and packages within Python.

Python Integration Guides

The following sections show how to create and use connections with the connector in common packages in Python:

• From SQLAlchemy
• From Pandas
• From Matplotlib
• From Petl

Complete List of Azure Cosmos DB Integration Quickstarts

The CData Python Connector for Azure Cosmos DB includes a Dialect class that enables integration with SQLAlchemy. Bear in mind that several aspects of connector functionality are not currently supported in SQLAlchemy 2.0 or above. If necessary, downgrade SQLAlchemy to version 1.4 or 1.3 before using this connector.

The following sections detail various aspects of this integration:

Connecting From SQLAlchemy

Reflecting Metadata With SQLAlchemy

Querying Data From SQLAlchemy

Modifying Data From SQLAlchemy

Connecting With a Dialect URL

from sqlalchemy import create_engine
engine = create_engine("cosmosdb:///?AccountEndpoint=myAccountEndpoint;AccountKey=myAccountKey;")

For SQLAlchemy 2.0, the dialect name is cosmosdb_2. To establish a connection, use the following URL format:

from sqlalchemy import create_engine
engine = create_engine("cosmosdb_2:///?AccountEndpoint=myAccountEndpoint;AccountKey=myAccountKey;")

SQLAlchemy can act as an Object-relational Map (ORM). This enables you to treat records of a database table as instantiable records. To leverage this functionality, you must reflect the underlying metadata in one of the following ways.

Note: The following examples employ SQLAlchemy 1.4.

Modeling Data Using a Mapping Class

from sqlalchemy.ext.declarative import declarative_base
Base = declarative_base()
class [CData].[Entities].Customers(Base):
	__tablename__ = "[CData].[Entities].Customers"
	_id = Column(String, primary_key=True)
	City = Column(String)
	CompanyName = Column(String)

Automatically Reflecting Metadata

from sqlalchemy import MetaData
from sqlalchemy.ext.automap import automap_base
meta = MetaData()
abase = automap_base(metadata=meta)
abase.prepare(autoload_with=engine)
[CData].[Entities].Customers = abase.classes.[CData].[Entities].Customers

You can also reflect a single table with an inspector. When reflecting this way, providing a list of specific columns to map is optional:

from sqlalchemy import MetaData, Table
from sqlalchemy import inspect
meta = MetaData()
insp = inspect(engine)
[CData].[Entities].Customers_table = Table("[CData].[Entities].Customers", meta)
insp.reflect_table([CData].[Entities].Customers_table, ["_id","CompanyName"])

After you use the steps in Connecting to connect, and use one of the methods in Reflecting Metadata to reflect some of the metadata, you can use a session object to query data.

Querying Data Using the Query Method

For example:

engine = create_engine("cosmosdb:///?AccountEndpoint=myAccountEndpoint;AccountKey=myAccountKey;")
factory = sessionmaker(bind=engine)
session = factory()
for instance in session.query([CData].[Entities].Customers).filter_by(Country="US"):
	print("_id: ", instance._id)
	print("City: ", instance.City)
	print("CompanyName: ", instance.CompanyName)
	print("---------")

Querying Data Using the Execute Method

[CData].[Entities].Customers_table = [CData].[Entities].Customers.metadata.tables["[CData].[Entities].Customers"]
for instance in session.execute([CData].[Entities].Customers_table.select().where([CData].[Entities].Customers_table.c.Country == "US")):
	print("Id: ", instance.Id)
	print("FullName: ", instance.Name)
	print("City: ", instance.BillingCity)
	print("---------")

Implicit Joining

from sqlalchemy.ext.declarative import declarative_base
from sqlalchemy import Column, String, Integer, DateTime, ForeignKey
from sqlalchemy.orm import sessionmaker, relationship

Base = declarative_base()
class Contact(Base):
	__tablename__ = "Contact"
	Id = Column(Integer, primary_key=True)
	Name = Column(String)
	Email = Column(String)
	BirthDate = Column(DateTime)
	AccountId = Column(String, ForeignKey("Account.Id"))
	Account_Link = relationship("Account", back_populates="Contact_Link")

class Account(Base):
	__tablename__ = "Account"
	Id = Column(String, primary_key=True)
	Name = Column(String)
	BillingCity = Column(String)
	NumberOfEmployees = Column(Integer)
	Contact_Link = relationship("Contact", order_by=Contact.Id, back_populates="Account_Link")

Once the relationship is established, the tables are queried simultaneously using the session's query() method. For example:

rs = session.query(Account, Contact).filter(Account.Id == Contact.AccountId)
for Ac, Ct in rs:
  print("AccountId: ", Ac.Id)
  print("AccountName: ", Ac.Name)
  print("ContactId: ", Ct.Id)
  print("ContactName: ", Ct.Name)

Other Join Forms

• Explicit condition (necessary if there are no foreign keys in your mapped classes):
  

  rs = session.query(Account, Contact).join(Contact, Account.Id == Contact.AccountId)
  for Ac, Ct in rs:

• Left-to-right relationship:
  

  rs = session.query(Account, Contact).join(Account.Contact_Link)
  for Ac, Ct in rs:

• Left-to-right relationship with explicit target:
  

  rs = session.query(Account, Contact).join(Contact, Account.Contact_Link)
  for Ac, Ct in rs:

• String form of a left-to-right relationship:
  

  rs = session.query(Account, Contact).join("Contact_Link")
  for Ac, Ct in rs:

SQLAlchemy ORM also exposes support for other clauses in SQL, such as ORDER BY, GROUP BY, LIMIT, and OFFSET. All of these are supported by this connector:

ORDER BY

rs = session.query([CData].[Entities].Customers).order_by([CData].[Entities].Customers.Balance)
for instance in rs:
	print("_id: ", instance._id)
	print("City: ", instance.City)
	print("CompanyName: ", instance.CompanyName)
	print("---------")

You can also use the session object's execute() method perform an ORDER BY. For example:

rs = session.execute([CData].[Entities].Customers_table.select().order_by([CData].[Entities].Customers_table.c.Balance))
for instance in rs:

GROUP BY

rs = session.query(func.count([CData].[Entities].Customers._id).label("CustomCount"), [CData].[Entities].Customers.City).group_by([CData].[Entities].Customers.City)
for instance in rs:
	print("Count: ", instance.CustomCount)
	print("City: ", instance.City)
	print("---------")

You can also use the session object's execute() method to perform a GROUP BY:

rs = session.execute([CData].[Entities].Customers_table.select().with_only_columns([func.count([CData].[Entities].Customers_table.c._id).label("CustomCount"), [CData].[Entities].Customers_table.c.City]).group_by([CData].[Entities].Customers_table.c.City))
for instance in rs:

LIMIT and OFFSET

rs = session.query([CData].[Entities].Customers).limit(25).offset(100)
for instance in rs:
	print("_id: ", instance._id)
	print("City: ", instance.City)
	print("CompanyName: ", instance.CompanyName)
	print("---------")

You can also use the session object's execute() method to set a LIMIT or OFFSET:

rs = session.execute([CData].[Entities].Customers_table.select().limit(25).offset(100))
for instance in rs:

Certain aggregate functions can also be used within SQLAlchemy by using the func module.

To import this module, execute:

from sqlalchemy.sql import func

Once func is imported, the following aggregate functions are available:

COUNT

rs = session.query(func.count([CData].[Entities].Customers._id).label("CustomCount"), [CData].[Entities].Customers.City).group_by([CData].[Entities].Customers.City)
for instance in rs:
	print("Count: ", instance.CustomCount)
	print("City: ", instance.City)
	print("---------")

You can also execute COUNT using the session object's execute() method:

rs = session.execute([CData].[Entities].Customers_table.select().with_only_columns([func.count([CData].[Entities].Customers_table.c._id).label("CustomCount"), [CData].[Entities].Customers_table.c.City])group_by([CData].[Entities].Customers_table.c.City))
for instance in rs:

SUM

This example calculates the cumulative amount of a numeric column in a set of groups.

rs = session.query(func.sum([CData].[Entities].Customers.Balance).label("CustomSum"), [CData].[Entities].Customers.City).group_by([CData].[Entities].Customers.City)
for instance in rs:
	print("Sum: ", instance.CustomSum)
	print("City: ", instance.City)
	print("---------")

You can also invoke SUM using the session object's execute() method.

rs = session.execute([CData].[Entities].Customers_table.select().with_only_columns([func.sum([CData].[Entities].Customers_table.c.Balance).label("CustomSum"), [CData].[Entities].Customers_table.c.City]).group_by([CData].[Entities].Customers_table.c.City))
for instance in rs:

AVG

rs = session.query(func.avg([CData].[Entities].Customers.Balance).label("CustomAvg"), [CData].[Entities].Customers.City).group_by([CData].[Entities].Customers.City)
for instance in rs:
	print("Avg: ", instance.CustomAvg)
	print("City: ", instance.City)
	print("---------")

You can also use the session object's execute() method to invoke AVG:

rs = session.execute([CData].[Entities].Customers_table.select().with_only_columns([func.avg([CData].[Entities].Customers_table.c.Balance).label("CustomAvg"), [CData].[Entities].Customers_table.c.City]).group_by([CData].[Entities].Customers_table.c.City))
for instance in rs:

MAX and MIN

rs = session.query(func.max([CData].[Entities].Customers.Balance).label("CustomMax"), func.min([CData].[Entities].Customers.Balance).label("CustomMin"), [CData].[Entities].Customers.City).group_by([CData].[Entities].Customers.City)
for instance in rs:
	print("Max: ", instance.CustomMax)
	print("Min: ", instance.CustomMin)
	print("City: ", instance.City)
	print("---------")

You can also use the session object's execute() method to invoke MAX and MIN:

rs = session.execute([CData].[Entities].Customers_table.select().with_only_columns([func.max([CData].[Entities].Customers_table.c.Balance).label("CustomMax"), func.min([CData].[Entities].Customers_table.c.Balance).label("CustomMin"), [CData].[Entities].Customers_table.c.City]).group_by([CData].[Entities].Customers_table.c.City))
for instance in rs:

Commands can be executed individually by the session with a call to "execute()".

Obtaining the Table Object

The query supplied to this method is constructed using the associated Table object of a mapped class. This Table object is obtained from the mapped class's metadata field, as below:

[CData].[Entities].Customers_table = [CData].[Entities].Customers.metadata.tables["[CData].[Entities].Customers"]

Once the table object is obtained, the write operations are executed in the following ways. The queries are executed immediately without the need for a call to "commit()":

Insert

The following example adds a new record to the table:

session.execute([CData].[Entities].Customers_table.insert(), {"City": "Jon Deere", "CompanyName": "Caterpillar"})

Update

The following example modifies an existing record in the table:

session.execute([CData].[Entities].Customers_table.update().where([CData].[Entities].Customers_table.c._id == "22").values(City="Jon Deere", CompanyName="Caterpillar"))

Delete

The following example removes an existing record from the table:

session.execute([CData].[Entities].Customers_table.delete().where([CData].[Entities].Customers_table.c._id == "22"))

When combined with the connector, Pandas can be used to generate data frames that contain your Azure Cosmos DB data. Once created, a data frame can be passed to various other Python packages.

Connecting

import pandas as pd
from sqlalchemy import create_engine
engine = create_engine("cosmosdb:///?AccountEndpoint=myAccountEndpoint;AccountKey=myAccountKey;")

Querying Data

df = pd.read_sql("""
	SELECT
	   City,
	   CompanyName,
     $exNumericCol;
	FROM [CData].[Entities].Customers;""", engine)
print(df)

Modifying Data

df = pd.DataFrame({"City": ["Jon Deere"], "CompanyName": ["Caterpillar"]})
df.to_sql("[CData].[Entities].Customers", con=engine, if_exists="append", index=False)

Matplotlib contains a number of tools that can graphically model Azure Cosmos DB data after being fed a data frame From Pandas.

Using PyPlot

from matplotlib import pyplot as plt

Once a Pandas data frame is obtained, it can be used to create a plot visualizing Azure Cosmos DB data. For example, the following plot generates and displays a bar graph relating City and Balance values:

df.plot(kind="bar", x="City", y=["Balance"])
plt.show()

The connector can be used to create ETL applications and pipelines for CSV data in Python using Petl.

Install Required Modules

pip install petl

Connecting

import petl as etl
import cdata.cosmosdb as mod
cnxn = mod.connect("AccountEndpoint=myAccountEndpoint;AccountKey=myAccountKey;")

Extract, Transform, and Load the Azure Cosmos DB Data

sql = "SELECT	City, CompanyName FROM [CData].[Entities].Customers "
table1 = etl.fromdb(cnxn,sql)

Loading Data

etl.tocsv(table1,'output.csv')

Modifying Data

table1 = [['City','CompanyName'],['Jon Deere','Caterpillar']]
etl.appenddb(table1,cnxn,'[CData].[Entities].Customers')

The extension supports schema discovery by using SQL queries to available System Tables.

Using SQL

• Tables and Views
• Columns
• Procedures

The connector possesses system tables that are used to discover the tables and views available in the data model. Of these system tables, "sys_tables" and "sys_views" are used to fetch information about the available tables and views respectively:

Tables

import cdata.cosmosdb as mod
conn = mod.connect("AccountEndpoint=myAccountEndpoint;AccountKey=myAccountKey;")
cur = conn.cursor()
cmd = "SELECT * FROM sys_tables"
cur.execute(cmd)
rs = cur.fetchall()
for row in rs:
	print(row)

Views

import cdata.cosmosdb as mod
conn = mod.connect("AccountEndpoint=myAccountEndpoint;AccountKey=myAccountKey;")
cur = conn.cursor()
cmd = "SELECT * FROM sys_views"
cur.execute(cmd, params)
rs = cur.fetchall()
for row in rs:
	print(row)

The available columns for any given table are fetched from a system table called "sys_tablecolumns". A specific table name is provided in the WHERE criteria to restrict the table from which the column information is fetched:

import cdata.cosmosdb as mod
conn = mod.connect("AccountEndpoint=myAccountEndpoint;AccountKey=myAccountKey;")
cur = conn.cursor()
cmd = "SELECT * FROM sys_tablecolumns WHERE TableName = '[CData].[Entities].Customers'"
cur.execute(cmd)
rs = cur.fetchall()
for row in rs:
	print(row)

Procedures

import cdata.cosmosdb as mod
conn = mod.connect("AccountEndpoint=myAccountEndpoint;AccountKey=myAccountKey;")
cur = conn.cursor()
cmd = "SELECT * FROM sys_procedures"
cur.execute(cmd)
rs = cur.fetchall()
for row in rs:
	print(row)

Parameters

import cdata.cosmosdb as mod
conn = mod.connect("AccountEndpoint=myAccountEndpoint;AccountKey=myAccountKey;")
cur = conn.cursor()
cmd = "SELECT * FROM sys_procedureparameters WHERE ProcedureName = 'EVAL'"
cur.execute(cmd)
rs = cur.fetchall()
for row in rs:
	print(row)

The CData Python Connector for Azure Cosmos DB supports several operations on data, including querying, deleting, modifying, and inserting.

SELECT Statements

See SELECT Statements for a syntax reference and examples.

INSERT Statements

See INSERT Statements for a syntax reference and examples.

UPDATE Statements

The primary key _id is required to update a record. See UPDATE Statements for a syntax reference and examples.

DELETE Statements

The primary key _id is required to delete a record. See DELETE Statements for a syntax reference and examples.

CACHE Statements

CACHE statements allow granular control over the connector's caching functionality. For a syntax reference and examples, see CACHE Statements.

For more information on the caching feature, see Caching Data.

EXECUTE Statements

Use EXECUTE or EXEC statements to execute stored procedures. See EXECUTE Statements for a syntax reference and examples.

Names and Quoting

• Table and column names are considered identifier names; as such, they are restricted to the following characters: [A-Z, a-z, 0-9, _:@].
• To use a table or column name with characters not listed above, the name must be quoted using square brackets ([name]) in any SQL statement.
• Parameter names can optionally start with the @ symbol (e.g., @p1 or @CustomerName) and cannot be quoted.
• Strings must be quoted using single quotes (e.g., 'John Doe').

The connector provides functions that are similar to those that are available with most standard databases. These functions are implemented in the CData provider engine and thus are available across all data sources with the same consistent API. Three categories of functions are available: string, date, and math.

The connector interprets all SQL function inputs as either strings or column identifiers, so you need to escape all literals as strings, with single quotes. For example, contrast the SQL Server syntax and connector syntax for the DATENAME function:

• SQL Server:
  

  SELECT DATENAME(yy,GETDATE())

• connector:
  

  SELECT DATENAME('yy',GETDATE())

String Functions

These functions perform string manipulations and return a string value. See STRING Functions for more details.

Date Functions

These functions perform date and date time manipulations. See DATE Functions for more details.

Math Functions

These functions provide mathematical operations. See MATH Functions for more details.

Projection Functions

These functions can be used to refine projections in your SQL query. See Projection Functions for more details.

Predicate Functions

These functions can be used to specify criteria in the WHERE clause of your SQL query. See Predicate Functions for more details.

ASCII(character_expression)

Returns the ASCII code value of the left-most character of the character expression.

• character_expression: The character expression.

                      SELECT ASCII('0');
                      --  Result: 48
                    

BASE64_ENCODE(input_binary)

Returns the Base64-encoded string form of a binary input.

• input_binary: The binary value to encode.

                        SELECT BASE64_ENCODE(BinaryData);
                    -- Result: 'QmFzZTY0RW5jb2RlZA=='
                    

BASE64_DECODE(input_string)

Returns the binary result of decoding a Base64-encoded string.

• input_string: The Base64-encoded string.

                        SELECT BASE64_DECODE('QmFzZTY0RW5jb2RlZA==');
                    -- Result: (binary output)
                    

CHAR(integer_expression)

Converts the integer ASCII code to the corresponding character.

• integer_expression: The integer from 0 through 255.

                      SELECT CHAR(48);
                      -- Result: '0'
                    

CHARINDEX(expressionToFind ,expressionToSearch [,start_location ])

Returns the starting position of the specified expression in the character string.

• expressionToFind: The character expression to find.
• expressionToSearch: The character expression, typically a column, to search.
• start_location: An optional character position to start searching for expressionToFind in expressionToSearch.

                      SELECT CHARINDEX('456', '0123456');
                      -- Result: 4

                      SELECT CHARINDEX('456', '0123456', 5);
                      -- Result: -1
                    

CHAR_LENGTH(character_expression),

Returns the number of UTF-8 characters present in the expression.

• character_expression: The set of characters to be evaluated for length.

				 SELECT CHAR_LENGTH('sample text') FROM Account LIMIT 1
				 -- Result: 11			
				

CONCAT(string_value1, string_value2, ..., string_valueN)

Returns the string that is the concatenation of two or more string values.

• string_value1: The first string to be concatenated.
• string_value2: The second string to be concatenated.
• string_valueN: (optional) Any additional strings to be concatenated.

                      SELECT CONCAT('Hello, ', 'world!');
                      -- Result: 'Hello, world!'
                    

CONTAINS(expressionToSearch, expressionToFind)

Returns 1 if expressionToFind is found within expressionToSearch; otherwise, 0.

• expressionToSearch: The character expression, typically a column, to search.
• expressionToFind: The character expression to find.

                      SELECT CONTAINS('0123456', '456');
                      -- Result: 1

                      SELECT CONTAINS('0123456', 'Not a number');
                      -- Result: 0
                    

ENDSWITH(character_expression, character_suffix)

Returns 1 if character_expression ends with character_suffix; otherwise, 0.

• character_expression: The character expression.
• character_suffix: The character suffix to search for.

                      SELECT ENDSWITH('0123456', '456');
                      -- Result: 1

                      SELECT ENDSWITH('0123456', '012');
                      -- Result: 0
                    

FILESIZE(uri)

Returns the number of bytes present in the file at the specified file path.

• uri: The path of the file from which to read the size.

				SELECT FILESIZE('C:/Users/User1/Desktop/myfile.txt');
				-- Result: 23684
				

FORMAT(value [, parseFormat], format )

Returns the value formatted with the specified format.

• value: The string to format.
• format: The string specifying the output syntax of the date or numeric format.
• parseFormat: The string specifying the input syntax of the date value. Not applicable to numeric types.

                      SELECT FORMAT(12.34, '#');
                      -- Result: 12

                      SELECT FORMAT(12.34, '#.###');
                      -- Result: 12.34

                      SELECT FORMAT(1234, '0.000E0');
                      -- Result: 1.234E3
                      
                      SELECT FORMAT('2019/01/01', 'yyyy-MM-dd');
                      -- Result: 2019-01-01
                      
                      SELECT FORMAT('20190101', 'yyyyMMdd', 'yyyy-MM-dd');
                      -- Result: '2019-01-01'
                    

HASHBYTES(algorithm, value)

Returns the hash of the input value as a byte array using the given algorithm. The supported algorithms are MD5, SHA1, SHA2_256, SHA2_512, SHA3_224, SHA3_256, SHA3_384, and SHA3_512.

• algorithm: The algorithm to use for hashing. Must be one of MD5, SHA1, SHA2_256, SHA2_512, SHA3_224, SHA3_256, SHA3_384, or SHA3_512.
• value: The value to hash. Must be either a string or byte array.

                      SELECT HASHBYTES('MD5', 'Test');
                      -- Result (byte array): 0x0CBC6611F5540BD0809A388DC95A615B
                    

INDEXOF(expressionToSearch, expressionToFind [,start_location ])

Returns the starting position of the specified expression in the character string.

• expressionToSearch: The character expression, typically a column, to search.
• expressionToFind: The character expression to find.
• start_location: An optional character position to start searching for expressionToFind in expressionToSearch.

                      SELECT INDEXOF('0123456', '456');
                      -- Result: 4

                      SELECT INDEXOF('0123456', '456', 5);
                      -- Result: -1
                    

ISALPHABETIC(character_expression)

Returns 1 if the character expression consists only of alphabetic characters; otherwise, 0.

• character_expression: The string expression to evaluate.

                      SELECT ISALPHABETIC('Hello');
                      -- Result: 1

                      SELECT ISALPHABETIC('Hello123');
                      -- Result: 0

                      SELECT ISALPHABETIC('Hello!');
                      -- Result: 0
                    

ISALPHANUMERIC(character_expression)

Returns 1 if the character expression consists only of alphabetic and numeric characters; otherwise, 0.

• character_expression: The string expression to evaluate.

                      SELECT ISALPHANUMERIC('Hello123');
                      -- Result: 1

                      SELECT ISALPHANUMERIC('123');
                      -- Result: 1

                      SELECT ISALPHANUMERIC('Hello.123');
                      -- Result: 0
                    

ISNUMERIC(character_expression)

Returns 1 if the character expression consists only of numeric digits and up to one decimal point; otherwise, 0.

• character_expression: The string expression to evaluate.

                      SELECT ISNUMERIC('123');
                      -- Result: 1

                      SELECT ISNUMERIC('123.45');
                      -- Result: 1

                      SELECT ISNUMERIC('123.45.67');
                      -- Result: 0

                      SELECT ISNUMERIC('12a3');
                      -- Result: 0
                    

JSON_EXTRACT(json, jsonpath)

Selects any value in a JSON array or object. The path to the array is specified in the jsonpath argument. Return value is numeric or null.

• json: The JSON document to extract.
• jsonpath: The XPath used to select the nodes. The JSONPath must be a string constant. The values of the nodes selected will be returned in a token-separated list.

                      SELECT JSON_EXTRACT('{"test": {"data": 1}}', '$.test');
                      -- Result: '{"data":1}'

                      SELECT JSON_EXTRACT('{"test": {"data": 1}}', '$.test.data');
                      -- Result: 1

                      SELECT JSON_EXTRACT('{"test": {"data": [1, 2, 3]}}', '$.test.data[1]');
                      -- Result: 2
                    

LEFT ( character_expression , integer_expression )

Returns the specified number of characters counting from the left of the specified string.

• character_expression: The character expression.
• integer_expression: The positive integer that specifies how many characters will be returned counting from the left of character_expression.

                      SELECT LEFT('1234567890', 3);
                      -- Result: '123'
                    

LEN(string_expression)

Returns the number of characters of the specified string expression.

• string_expression: The string expression.

                      SELECT LEN('12345');
                      -- Result: 5
                    

LOCATE(substring,string)

Returns an integer representing how many characters into the string the substring appears.

• substring: The substring to find inside larger string.
• string: The larger string that is searched for the substring.
• start locations: An optional integer that sets the character position (offset) from which to start searching.

				SELECT LOCATE('sample','XXXXXsampleXXXXX');
				-- Result: 6

                SELECT LOCATE('sample', 'XXXXXsampleXXXXX', 7)
                -- Result: 0
				

LOWER ( character_expression )

Returns the character expression with the uppercase character data converted to lowercase.

• character_expression: The character expression.

                      SELECT LOWER('MIXED case');
                      -- Result: 'mixed case'
                    

LTRIM(character_expression)

Returns the character expression with leading blanks removed.

• character_expression: The character expression.

                      SELECT LTRIM('     trimmed');
                      -- Result: 'trimmed'
                    

MASK(string_expression, mask_character [, start_index [, end_index ]])

Replaces the characters between start_index and end_index with the mask_character within the string.

• string_expression: The string expression to be searched.
• mask_character: The character to mask with.
• start_index: The optional number of characters to leave unmasked at beginning of string. Defaults to 0.
• end_index: The optional number of characters to leave unmasked at end of string. Defaults to 0.

                        SELECT MASK('1234567890','*',);
                        -- Result: '**********'
                        SELECT MASK('1234567890','*', 4);
                        -- Result: '1234******'
                        SELECT MASK('1234567890','*', 4, 2);
                        -- Result: '1234****90'  
                    

NCHAR(integer_expression)

Returns the Unicode character with the specified integer code as defined by the Unicode standard.

• integer_expression: The integer from 0 through 65535 (0 through xFFFF).

OCTET_LENGTH(character_expression),

Returns the number of bytes present in the expression.

• character_expression: The set of characters to be be evaluated.

				 SELECT OCTET_LENGTH('text') FROM Account LIMIT 1
				 -- Result: 4
				

PATINDEX(pattern, expression)

Returns the starting position of the first occurrence of the pattern in the expression. Returns 0 if the pattern is not found.

• pattern: The character expression that contains the sequence to be found. The wild-card character % can be used only at the start or end of the expression.
• expression: The expression, typically a column, to search for the pattern.

                      SELECT PATINDEX('123%', '1234567890');
                      -- Result: 1

                      SELECT PATINDEX('%890', '1234567890');
                      -- Result: 8

                      SELECT PATINDEX('%456%', '1234567890');
                      -- Result: 4
                    

POSITION(expressionToFind IN expressionToSearch)

Returns the starting position of the specified expression in the character string.

• expressionToFind: The character expression to find.
• expressionToSearch: The character expression, typically a column, to search.

                      SELECT POSITION('456' IN '123456');
                      -- Result: 4

                      SELECT POSITION('x' IN '123456');
                      -- Result: 0
                    

QUOTENAME(character_string [, quote_character])

Returns a valid SQL Server-delimited identifier by adding the necessary delimiters to the specified Unicode string.

• character_string: The string of Unicode character data. The string is limited to 128 characters. Inputs greater than 128 characters return null.
• quote_character: An optional single character to be used as the delimiter. These include:
  • a single quotation mark (')
  • a left or right bracket ([])
  • a double quotation mark (")
  • a left or right parenthesis ( () )
  • a greater or less than sign (><)
  • a left or right brace ({})
  • a backtick (`)

  If quote_character is not specified brackets are used. If an unacceptable character is supplied, it returns NULL.

                      SELECT QUOTENAME('table_name');
                      -- Result: '[table_name]'

                      SELECT QUOTENAME('table_name', '"');
                      -- Result: '"table_name"'

                      SELECT QUOTENAME('table_name', '[');
                      -- Result: '[table_name]'
                    

REGEXP_REPLACE(expr, pattern [, replacement [, position [, occurrence [, match_type]]]])

Replaces occurrences of a regular expression pattern in the input string with a specified value and returns the resulting string.

• expr: The string expression to be searched.
• pattern: The regular expression pattern to match.
• replacement: (optional) The string to replace each matched occurrence of pattern with. Supports backreferences \1 through \9 and escape sequences \n, \r, \t, and \\. By default, this argument is an empty string, meaning matched portions are removed from the output string.
• position: (optional) The 1-based starting position used when searching for regular expression matches in expr. The default is 1. All characters prior to the starting position are included in the output string unaltered. Skipped characters are ignored when calculating regular expression matches, even if they match pattern.
• occurrence: (optional) Specifies whether all occurrences of pattern,, or only a specific occurrence of pattern are replaced. The default is 0, which means all occurrences of pattern are replaced with replacement. Set to 1 to only replace the first instance of the pattern; 2 to replace the second; etc.
• match_type: (optional) Modifiers used to customize matching behavior. Supported values are: 'c' (case-sensitive, default), 'i' (case-insensitive), 'm' (multiline), 'n' (dot matches newline), 'x' (extended mode). These can be freely combined by including the letters back to back. For example, 'im' applies the functionality of both 'i' and 'm'. The regular expression syntax used is that of the Extended mode ('x') ignores whitespace and allows inline comments. If the pattern needs to match a literal space, it must be explicitly escaped.

                      SELECT REGEXP_REPLACE('abc123def456', '\d+', 'NUM');
                      -- Result: 'abcNUMdefNUM'

                      SELECT REGEXP_REPLACE('Hello\nHELLO\nhello', '^hello', 'X', 1, 0, 'im');
                      -- Result: 'X\nX\nX'
                    

REPLACE(string_expression, string_pattern, string_replacement)

Replaces all occurrences of a string with another string.

• string_expression: The string expression to be searched. This can be a character or binary data type.
• string_pattern: The substring to be found. Cannot be an empty string.
• string_replacement: The replacement string.

                      SELECT REPLACE('1234567890', '456', '|');
                      -- Result: '123|7890'

                      SELECT REPLACE('123123123', '123', '.');
                      -- Result: '...'

                      SELECT REPLACE('1234567890', 'a', 'b');
                      -- Result: '1234567890'
                    

REPLICATE ( string_expression ,integer_expression )

Repeats the string value the specified number of times.

• string_expression: The string to replicate.
• integer_expression: The repeat count.

                      SELECT REPLACE('x', 5);
                      -- Result: 'xxxxx'
                    

REVERSE ( string_expression )

Returns the reverse order of the string expression.

• string_expression: The string.

                      SELECT REVERSE('1234567890');
                      -- Result: '0987654321'
                    

RIGHT ( character_expression , integer_expression )

Returns the right part of the string with the specified number of characters.

• character_expression: The character expression.
• integer_expression: The positive integer that specifies how many characters of the character expression will be returned.

                      SELECT RIGHT('1234567890', 3);
                      -- Result: '890'
                    

RTRIM(character_expression)

Returns the character expression after it removes trailing blanks.

• character_expression: The character expression.

                      SELECT RTRIM('trimmed     ');
                      -- Result: 'trimmed'
                    

SOUNDEX(character_expression)

Returns the four-character Soundex code, based on how the string sounds when spoken.

• character_expression: The alphanumeric expression of character data.

                      SELECT SOUNDEX('smith');
                      -- Result: 'S530'
                    

SPACE(repeatcount)

Returns the string that consists of repeated spaces.

• repeatcount: The number of spaces.

                      SELECT SPACE(5);
                      -- Result: '     '
                    

SPLIT(string, delimiter, offset)

Returns a section of the string between to delimiters.

• string: The string to split.
• delimiter: The character to split the string with.
• offset: The number of the split to return. Positive numbers are treated as offsets from the left, and negative numbers are treated as offsets from the right.

                      SELECT SPLIT('a/b/c/d', '/', 1);
                      -- Result: 'a'
                      SELECT SPLIT('a/b/c/d', '/', -2);
                      -- Result: 'c'
                    

STARTSWITH(character_expression, character_prefix)

Returns 1 if character_expression starts with character_prefix; otherwise, 0.

• character_expression: The character expression.
• character_prefix: The character prefix to search for.

                      SELECT STARTSWITH('0123456', '012');
                      -- Result: 1

                      SELECT STARTSWITH('0123456', '456');
                      -- Result: 0
                    

STR ( float_expression [ , integer_length [ , integer_decimal ] ] )

Returns the character data converted from the numeric data. For example, STR(123.45, 6, 1) returns 123.5.

• float_expression: The float expression.
• length: The optional total length to return. This includes decimal point, sign, digits, and spaces. The default is 10.
• decimal: The optional number of places to the right of the decimal point. The decimal must be less than or equal to 16.

                      SELECT STR('123.456');
                      -- Result: '123'

                      SELECT STR('123.456', 2);
                      -- Result: '**'

                      SELECT STR('123.456', 10, 2);
                      -- Result: '123.46'
                    

STUFF(character_expression , integer_start , integer_length , replaceWith_expression)

Inserts a string into another string. It deletes the specified length of characters in the first string at the start position and then inserts the second string into the first string at the start position.

• character_expression: The string expression.
• start: The integer value that specifies the location to start deletion and insertion. If start or length is negative, null is returned. If start is longer than the string to be modified, character_expression, null is returned.
• length: The integer that specifies the number of characters to delete. If length is longer than character_expression, deletion occurs up to the last character in replaceWith_expression.
• replaceWith_expression: The expression of character data that will replace length characters of character_expression beginning at the start value.

                      SELECT STUFF('1234567890', 3, 2, 'xx');
                      -- Result: '12xx567890'
                    

SUBSTRING(string_value FROM start FOR length)

Returns the part of the string with the specified length; starts at the specified index.

• string_value: The character string.
• start: The positive integer that specifies the start index of characters to return.
• length: Optional. The positive integer that specifies how many characters will be returned.

                      SELECT SUBSTRING('1234567890' FROM 3 FOR 2);
                      -- Result: '34'

                      SELECT SUBSTRING('1234567890' FROM 3);
                      -- Result: '34567890'
                    

                    SELECT SUBSTRING('1234567890', 3, 2)
                    --Result: '34'
                    SELECT SUBSTRING('1234567890', 3)
                    --Result: '34567890'
                    

TEXT_ENCODE(input_string, charset)

Returns binary output by encoding a string using the specified character set.

• input_string: The plain text string.
• charset: The character set to use, such as 'UTF-8', 'ISO-8859-1'.

                    SELECT TEXT_ENCODE('Café', 'UTF-8');
                    -- Result: (binary output)
                    

TEXT_DECODE(input_binary, charset)

Returns a string decoded from binary data using the specified character set.

• input_binary: The binary value to decode.
• charset: The character set used for decoding.

                    SELECT TEXT_DECODE(BinaryData, 'UTF-8');
                    -- Result: 'Café'
                    

TOSTRING(string_value1)

Converts the value of this instance to its equivalent string representation.

• string_value1: The string to be converted.

                      SELECT TOSTRING(123);
                      -- Result: '123'

                      SELECT TOSTRING(123.456);
                      -- Result: '123.456'

                      SELECT TOSTRING(null);
                      -- Result: ''
                    

TRIM(trimspec trimchar FROM string_value)

Returns the character expression with leading and/or trailing blanks removed.

• trimspec: Optional. If included must be one of the keywords BOTH, LEADING or TRAILING.
• trimchar: Optional. If included should be a one-character string value.
• string_value: The string value to trim.

                      SELECT TRIM('     trimmed     ');
                      -- Result: 'trimmed'

                      SELECT TRIM(LEADING FROM '     trimmed     ');
                      -- Result: 'trimmed     '

                      SELECT TRIM('-' FROM '-----trimmed-----');
                      -- Result: 'trimmed'

                      SELECT TRIM(BOTH '-' FROM '-----trimmed-----');
                      -- Result: 'trimmed'

                      SELECT TRIM(TRAILING '-' FROM '-----trimmed-----');
                      -- Result: '-----trimmed'
                    

UNICODE(ncharacter_expression)

Returns the integer value defined by the Unicode standard of the first character of the input expression.

• ncharacter_expression: The Unicode character expression.

UPPER ( character_expression )

Returns the character expression with lowercase character data converted to uppercase.

• character_expression: The character expression.

                      SELECT UPPER('MIXED case');
                      -- Result: 'MIXED CASE'
                    

XML_EXTRACT(xml, xpath [, separator])

Extracts an XML document using the specified XPath to flatten the XML. A comma is used to separate the outputs by default, but this can be changed by specifying the third parameter.

• xml: The XML document to extract.
• xpath: The XPath used to select the nodes. The nodes selected will be returned in a token-separated list.
• separator: The optional token used to separate the items in the flattened response. If this is not specified, the separator will be a comma.

                      SELECT XML_EXTRACT('<vowels><ch>a</ch><ch>e</ch><ch>i</ch><ch>o</ch><ch>u</ch></vowels>', '/vowels/ch');
                      -- Result: 'a,e,i,o,u'

                      SELECT XML_EXTRACT('<vowels><ch>a</ch><ch>e</ch><ch>i</ch><ch>o</ch><ch>u</ch></vowels>', '/vowels/ch', ';');
                      -- Result: 'a;e;i;o;u'
                    

ABS ( numeric_expression )

Returns the absolute (positive) value of the specified numeric expression.

• numeric_expression: The expression of an indeterminate numeric data type except for the bit data type.

                      SELECT ABS(15);
                      -- Result: 15

                      SELECT ABS(-15);
                      -- Result: 15
                    

ACOS ( float_expression )

Returns the arc cosine, the angle in radians whose cosine is the specified float expression.

• float_expression: The float expression that specifies the cosine of the angle to be returned. Values outside the range from -1 to 1 return null.

                      SELECT ACOS(0.5);
                      -- Result: 1.0471975511966
                    

ASIN ( float_expression )

Returns the arc sine, the angle in radians whose sine is the specified float expression.

• float_expression: The float expression that specifies the sine of the angle to be returned. Values outside the range from -1 to 1 return null.

                      SELECT ASIN(0.5);
                      -- Result: 0.523598775598299
                    

ATAN ( float_expression )

Returns the arc tangent, the angle in radians whose tangent is the specified float expression.

• float_expression: The float expression that specifies the tangent of the angle to be returned.

                      SELECT ATAN(10);
                      -- Result: 1.47112767430373
                    

ATN2 ( float_expression1 , float_expression2 )

Returns the angle in radians between the positive x-axis and the ray from the origin to the point (y, x) where x and y are the values of the two specified float expressions.

• float_expression1: The float expression that is the y-coordinate.
• float_expression2: The float expression that is the x-coordinate.

                      SELECT ATN2(1, 1);
                      -- Result: 0.785398163397448
                    

CEILING ( numeric_expression ) or CEIL( numeric_expression )

Returns the smallest integer greater than or equal to the specified numeric expression.

• numeric_expression: The expression of an indeterminate numeric data type except for the bit data type.

                      SELECT CEILING(1.3);
                      -- Result: 2

                      SELECT CEILING(1.5);
                      -- Result: 2

                      SELECT CEILING(1.7);
                      -- Result: 2
                    

COS ( float_expression )

Returns the trigonometric cosine of the specified angle in radians in the specified expression.

• float_expression: The float expression of the specified angle in radians.

                      SELECT COS(1);
                      -- Result: 0.54030230586814
                    

COT ( float_expression )

Returns the trigonometric cotangent of the angle in radians specified by float_expression.

• float_expression: The float expression of the angle in radians.

                      SELECT COT(1);
                      -- Result: 0.642092615934331
                    

DEGREES ( numeric_expression )

Returns the angle in degrees for the angle specified in radians.

• numeric_expression: The angle in radians, an expression of an indeterminate numeric data type except for the bit data type.

                      SELECT DEGREES(3.1415926);
                      -- Result: 179.999996929531
                    

EXP ( float_expression )

Returns the exponential value of the specified float expression. For example, EXP(LOG(20)) is 20.

• float_expression: The float expression.

                      SELECT EXP(2);
                      -- Result: 7.38905609893065
                    

EXPR ( expression )

Evaluates the expression.

• expression: The expression. Operators allowed are +, -, *, /, ==, !=, >, <, >=, and <=.

                      SELECT EXPR('1 + 2 * 3');
                      -- Result: 7

                      SELECT EXPR('1 + 2 * 3 == 7');
                      -- Result: true
                    

FLOOR ( numeric_expression )

Returns the largest integer less than or equal to the numeric expression.

• numeric_expression: The expression of an indeterminate numeric data type except for the bit data type.

                      SELECT FLOOR(1.3);
                      -- Result: 1

                      SELECT FLOOR(1.5);
                      -- Result: 1

                      SELECT FLOOR(1.7);
                      -- Result: 1
                    

GREATEST(int1,int2,....)

Returns the greatest of the supplied integers.

				SELECT GREATEST(3,5,8,10,1)
				-- Result: 10			
				

HEX(value)

Returns a the equivalent hex for the input value.

• value: A string or numerical value to be converted into hex.

				SELECT HEX(866849198);
				-- Result: 33AB11AE
				
				SELECT HEX('Sample Text');
				-- Result: 53616D706C652054657874
				

JSON_AVG(json, jsonpath)

Computes the average value of a JSON array within a JSON object. The path to the array is specified in the jsonpath argument. Return value is numeric or null.

• json: The JSON document to compute.
• jsonpath: The JSONPath used to select the nodes. [x], [2..], [..8], or [1..12] are accepted. [x] selects all nodes.

                      SELECT JSON_AVG('[1,2,3,4,5]', '$[x]');
                      -- Result: 3

                      SELECT JSON_AVG('{"test": {"data": [1,2,3,4,5]}}', '$.test.data[x]');
                      -- Result: 3

                      SELECT JSON_AVG('{"test": {"data": [1,2,3,4,5]}}', '$.test.data[3..]');
                      -- Result: 4.5
                    

JSON_COUNT(json, jsonpath)

Returns the number of elements in a JSON array within a JSON object. The path to the array is specified in the jsonpath argument. Return value is numeric or null.

• json: The JSON document to compute.
• jsonpath: The JSONPath used to select the nodes. [x], [2..], [..8], or [1..12] are accepted. [x] selects all nodes.

                      SELECT JSON_COUNT('[1,2,3,4,5]', '$[x]');
                      -- Result: 5

                      SELECT JSON_COUNT('{"test": {"data": [1,2,3,4,5]}}', '$.test.data[x]');
                      -- Result: 5

                      SELECT JSON_COUNT('{"test": {"data": [1,2,3,4,5]}}', '$.test.data[3..]');
                      -- Result: 2
                    

JSON_MAX(json, jsonpath)

Gets the maximum value in a JSON array within a JSON object. The path to the array is specified in the jsonpath argument. Return value is numeric or null.

• json: The JSON document to compute.
• jsonpath: The JSONPath used to select the nodes. [x], [2..], [..8], or [1..12] are accepted. [x] selects all nodes.

                      SELECT JSON_MAX('[1,2,3,4,5]', '$[x]');
                      -- Result: 5

                      SELECT JSON_MAX('{"test": {"data": [1,2,3,4,5]}}', '$.test.data[x]');
                      -- Result: 5

                      SELECT JSON_MAX('{"test": {"data": [1,2,3,4,5]}}', '$.test.data[..3]');
                      -- Result: 4
                    

JSON_MIN(json, jsonpath)

Gets the minimum value in a JSON array within a JSON object. The path to the array is specified in the jsonpath argument. Return value is numeric or null.

• json: The JSON document to compute.
• jsonpath: The JSONPath used to select the nodes. [x], [2..], [..8], or [1..12] are accepted. [x] selects all nodes.

                      SELECT JSON_MIN('[1,2,3,4,5]', '$[x]');
                      -- Result: 1

                      SELECT JSON_MIN('{"test": {"data": [1,2,3,4,5]}}', '$.test.data[x]');
                      -- Result: 1

                      SELECT JSON_MIN('{"test": {"data": [1,2,3,4,5]}}', '$.test.data[3..]');
                      -- Result: 4
                    

JSON_SUM(json, jsonpath)

Computes the summary value in JSON according to the JSONPath expression. Return value is numeric or null.

• json: The JSON document to compute.
• jsonpath: The JSONPath used to select the nodes. [x], [2..], [..8], or [1..12] are accepted. [x] selects all nodes.

                      SELECT JSON_SUM('[1,2,3,4,5]', '$[x]');
                      -- Result: 15

                      SELECT JSON_SUM('{"test": {"data": [1,2,3,4,5]}}', '$.test.data[x]');
                      -- Result: 15

                      SELECT JSON_SUM('{"test": {"data": [1,2,3,4,5]}}', '$.test.data[3..]');
                      -- Result: 9
                    

LEAST(int1,int2,....)

Returns the least of the supplied integers.

				SELECT LEAST(3,5,8,10,1)
				-- Result: 1			
				

LOG ( float_expression [, base ] )

Returns the natural logarithm of the specified float expression.

• float_expression: The float expression.
• base: The optional integer argument that sets the base for the logarithm.

                      SELECT LOG(7.3890560);
                      -- Result: 1.99999998661119
                    

LOG10 ( float_expression )

Returns the base-10 logarithm of the specified float expression.

• float_expression: The expression of type float.

                      SELECT LOG10(10000);
                      -- Result: 4
                    

MOD(dividend,divisor)

Returns the integer value associated with the remainder when dividing the dividend by the divisor.

• dividend: The number to take the modulus of.
• divisor: The number to divide the dividend by when determining the modulus.

				SELECT MOD(10,3);
				-- Result: 1
				

NEGATE(real_number)

Returns the opposite to the real number input.

• real_number: The real number to find the opposite of.

				SELECT NEGATE(10);
				-- Result: -10
				
				SELECT NEGATE(-12.4)
				--Result: 12.4
				

PI ( )

Returns the constant value of pi.

                  SELECT PI()
                  -- Result: 3.14159265358979 
                

POWER ( float_expression , y )

Returns the value of the specified expression raised to the specified power.

• float_expression: The float expression.
• y: The power to raise float_expression to.

                      SELECT POWER(2, 10);
                      -- Result: 1024

                      SELECT POWER(2, -2);
                      -- Result: 0.25
                    

RADIANS ( float_expression )

Returns the angle in radians of the angle in degrees.

• float_expression: The degrees of the angle as a float expression.

                      SELECT RADIANS(180);
                      -- Result: 3.14159265358979
                    

RAND ( [ integer_seed ] )

Returns a pseudorandom float value from 0 through 1, exclusive.

• seed: The optional integer expression that specifies the seed value. If seed is not specified, a seed value at random will be assigned.

                      SELECT RAND();
                      -- This result may be different, since the seed is randomized
                      -- Result: 0.873159630165044

                      SELECT RAND(1);
                      -- This result will always be the same, since the seed is constant
                      -- Result: 0.248668584157093
                    

ROUND ( numeric_expression [ ,integer_length] [ ,function ] )

Returns the numeric value rounded to the specified length or precision.

• numeric_expression: The expression of a numeric data type.
• length: The optional precision to round the numeric expression to. When this is omitted, the default behavior will be to round to the nearest whole number.
• function: The optional type of operation to perform. When the function parameter is omitted or has a value of 0 (default), numeric_expression is rounded. When a value other than 0 is specified, numeric_expression is truncated.

                      SELECT ROUND(1.3, 0);
                      -- Result: 1

                      SELECT ROUND(1.55, 1);
                      -- Result: 1.6

                      SELECT ROUND(1.7, 0, 0);
                      -- Result: 2

                      SELECT ROUND(1.7, 0, 1);
                      -- Result: 1
                      
                      SELECT ROUND (1.24);
                      -- Result: 1.0
                    

SIGN ( numeric_expression )

Returns the positive sign (1), 0, or negative sign (-1) of the specified expression.

• numeric_expression: The expression of an indeterminate data type except for the bit data type.

                      SELECT SIGN(0);
                      -- Result: 0

                      SELECT SIGN(10);
                      -- Result: 1

                      SELECT SIGN(-10);
                      -- Result: -1
                    

SIN ( float_expression )

Returns the trigonometric sine of the angle in radians.

• float_expression: The float expression specifying the angle in radians.

                     SELECT SIN(1);
                     -- Result: 0.841470984807897
                    

SQRT ( float_expression )

Returns the square root of the specified float value.

• float_expression: The expression of type float.

                      SELECT SQRT(100);
                      -- Result: 10
                    

SQUARE ( float_expression )

Returns the square of the specified float value.

• float_expression: The expression of type float.

                      SELECT SQUARE(10);
                      -- Result: 100

                      SELECT SQUARE(-10);
                      -- Result: 100
                    

TAN ( float_expression )

Returns the tangent of the input expression.

• float_expression: The expression of type float.

                      SELECT TAN(1);
                      -- Result: 1.5574077246549
                    

TRUNC(decimal_number,precision)

Returns the supplied decimal number truncated to have the supplied decimal precision.

• decimal_number: The decimal value to truncate.
• precision: The number of decimal places to truncate the decimal number to.

				SELECT TRUNC(10.3423,2);
				-- Result: 10.34
				

_ROW_NUMBER_()

Returns a row index as an additional column.

				SELECT ColumnName, _ROW_NUMBER_() FROM TableName
				-- Result: ColumnData, 0
				ColumnData2, 1
				ColumnData3, 2
				

CURRENT_DATE()

Returns the current date value.

                  SELECT CURRENT_DATE();
                  -- Result: 2018-02-01
                

CURRENT_TIMESTAMP()

Returns the current time stamp of the database system as a datetime value. This value is equal to GETDATE and SYSDATETIME, and is always in the local timezone.

                  SELECT CURRENT_TIMESTAMP();
                  -- Result: 2018-02-01 03:04:05
                

DATEADD (datepart , integer_number , date [, dateformat])

Returns the datetime value that results from adding the specified number (a signed integer) to the specified date part of the date.

• datepart: The part of the date to add the specified number to. The valid values and abbreviations are
  • year (yy, yyyy)
  • quarter (qq, q)
  • month (mm, m)
  • week (wk, ww)
  • weekday (dw)
  • dayofyear (dy, y)
  • day (dd, d)
  • hour (hh)
  • minute (mi, n)
  • second (ss, s)
  • millisecond (ms)
• number: The number to be added.
• date: The expression of the datetime data type.
• dateformat: The optional output date format.

                  SELECT DATEADD('d', 5, '2018-02-01');
                  -- Result: 2018-02-06

                  SELECT DATEADD('hh', 5, '2018-02-01 00:00:00');
                  -- Result: 2018-02-01 05:00:00
                

DATEDIFF ( datepart , startdate , enddate )

Returns the difference (a signed integer) of the specified time interval between the specified start date and end date.

• datepart: The part of the date that is the time interval of the difference between the start date and end date. The valid values and abbreviations are:
  • Year (year, yyyy, yy)
  • Quarter (quarter, qq, q)
  • Month (month, mm, m)
  • Week (week, wk, ww)
  • Weekday (weekday, dw)
  • Dayofyear (dayofyear, dy, y)
  • Day (day, dd, d)
  • Hour (hour, hh)
  • Minute (minute, mi, n)
  • Second (second, ss, s)
  • Millisecond (millisecond, ms)
• startdate: The datetime expression of the start date.
• enddate: The datetime expression of the end date.

                  SELECT DATEDIFF('d', '2018-02-01', '2018-02-10');
                  -- Result: 9

                  SELECT DATEDIFF('hh', '2018-02-01 00:00:00', '2018-02-01 12:00:00');
                  -- Result: 12
                

DATE_FORMAT(date,format)

Returns the date or timestamp in the format specified. This function mirrors the MySQL DATE_FORMAT function.

• date: A date or timestamp string.
• format: The specifier string of the desired output format. The list of supported format specifiers comes from the MySQL DATE_FORMAT function (see link to MySQL documentation above).

					SELECT DATE_FORMAT('9/4/2021 3:11:53 AM','%h')
					-- Result: 03
				  

DATEFROMPARTS(integer_year, integer_month, integer_day)

Returns the datetime value for the specified year, month, and day.

• year: The integer expression specifying the year.
• month: The integer expression specifying the month.
• day: The integer expression specifying the day.

                    SELECT DATEFROMPARTS(2018, 2, 1);
                    -- Result: 2018-02-01
                  

DATENAME(datepart , date)

Returns the character string that represents the specified date part of the specified date.

• datepart: The part of the date to return. The valid values and abbreviations are year (yy, yyyy), quarter (qq, q), month (mm, m), dayofyear (dy, y), day (dd, d), week (wk, ww), weekday (dw), hour (hh), minute (mi, n), second (ss, s), millisecond (ms), microsecond (mcs), and nanosecond (ns).
• date: The datetime expression.

                     SELECT DATENAME('yy', '2018-02-01');
                     -- Result: '2018'

                     SELECT DATENAME('dw', '2018-02-01');
                     -- Result: 'Thursday'
                   

DATEPART(datepart, date [,integer_datefirst])

Returns a character string that represents the specified date part of the specified date.

• datepart: The part of the date to return. The valid values and abbreviations are year (yy, yyyy), quarter (qq, q), month (mm, m), dayofyear (dy, y), day (dd, d), week (wk, ww), weekday (dw), hour (hh), minute (mi, n), second (ss, s), millisecond (ms), microsecond (mcs), nanosecond (ns), ISODOW, ISO_WEEK (isoweek, isowk,isoww), and ISOYEAR.
• date: The datetime string.
• datefirst: The optional integer representing the first day of the week. The default is 7, Sunday.

                    SELECT DATEPART('yy', '2018-02-01');
                    -- Result: 2018

                    SELECT DATEPART('dw', '2018-02-01');
                    -- Result: 5
                  

DATETIMEFROMPARTS(integer_year, integer_month, integer_day, integer_hour, integer_minute, integer_seconds, integer_milliseconds)

Returns the datetime value for the specified date parts.

• year: The integer expression specifying the year.
• month: The integer expression specifying the month.
• day: The integer expression specifying the day.
• hour: The integer expression specifying the hour.
• minute: The integer expression specifying the minute.
• seconds: The integer expression specifying the seconds.
• milliseconds: The integer expression specifying the milliseconds.

                    SELECT DATETIMEFROMPARTS(2018, 2, 1, 1, 2, 3, 456);
                    -- Result: 2018-02-01 01:02:03.456
                  

DATETIME2FROMPARTS(integer_year, integer_month, integer_day, integer_hour, integer_minute, integer_seconds, integer_fractions, integer_precision)

Returns the datetime value for the specified date parts.

• year: The integer expression specifying the year.
• month: The integer expression specifying the month.
• day: The integer expression specifying the day.
• hour: The integer expression specifying the hour.
• minute: The integer expression specifying the minute.
• seconds: The integer expression specifying the seconds.
• fractions: The integer expression specifying the fractions of the second.
• precision: The integer expression specifying the precision of the fraction.

				    SELECT DATETIME2FROMPARTS(2018, 2, 1, 1, 2, 3, 456, 3);
                    -- Result: 2018-02-01 01:02:03.456
                  

DATE_TRUNC(date, datepart)

Truncates the date to the precision of the given date part. Modeled after the Oracle TRUNC function.

• date: The datetime string that specifies the date.
• datepart: Refer to the Oracle documentation for valid datepart syntax.

				    SELECT DATE_TRUNC('05-04-2005', 'YY');
                    -- Result: '1/1/2005'
					
                    SELECT DATE_TRUNC('05-04-2005', 'MM');
                    -- Result: '5/1/2005'                    
                  

DATE_TRUNC2(datepart, date, [weekday])

Truncates the date to the precision of the given date part. Modeled after the PostgreSQL date_trunc function.

• datepart: One of 'millennium', 'century', 'decade', 'year', 'quarter', 'month', 'week', 'day', 'hour', 'minute' or 'second'.
• date: The datetime string that specifies the date.
• weekday: The optional day of the week to use as the first day for 'week'. One of 'sunday', 'monday', etc.

                    SELECT DATE_TRUNC2('year', '2020-02-04');
                    -- Result: '2020-01-01'

                    SELECT DATE_TRUNC2('week', '2020-02-04', 'monday');
                    -- Result: '2020-02-02', which is the previous Monday
                  

DAY(date)

Returns the integer that specifies the day component of the specified date.

• date: The datetime string that specifies the date.

                    SELECT DAY('2018-02-01');
                    -- Result: 1
                  

DAYNAME(date)

Returns the name of the day of the week of the specified date.

• date: The datetime string that specifies the date.

                    SELECT DAYNAME('8/18/2021');
                    -- Result: Wednesday
                  

DAYOFMONTH(date)

• date: The datetime string that specifies the date.

				  SELECT DAYOFMONTH('04/15/2000');
				  -- Result: 15
				  

DAYOFWEEK(date)

• date: The datetime string that specifies the date.

				  SELECT DAYOFWEEK('04/15/2000');
				  -- Result: 7
				  

DAYOFYEAR(date)

• date: The datetime string that specifies the date.

				  SELECT DAYOFYEAR('04/15/2000');
				  -- Result: 106
				  

EOMONTH(date [, integer_month_to_add ]) or LAST_DAY(date)

Returns the last day of the month that contains the specified date with an optional offset.

• date: The datetime expression specifying the date for which to return the last day of the month.
• integer_month_to_add: The optional integer expression specifying the number of months to add to the date before calculating the end of the month.

                  SELECT EOMONTH('2018-02-01');
                  -- Result: 2018-02-28
                  
                  SELECT LAST_DAY('2018-02-01');
                  -- Result: 2018-02-28

                  SELECT EOMONTH('2018-02-01', 2);
                  -- Result: 2018-04-30
                

EXTRACT(date_part FROM date_column_name)

Returns the last day of the month that contains the specified date with an optional offset.

• date_part: One of the following date components: YEAR, MONTH, DAY, HOUR, MINUTE, SECOND.
• date_column_name: The name of a date column in a table.

                  SELECT EXTRACT(YEAR FROM DateColumn)
                  -- Result: 2021
                

FDWEEK(date)

• date: The datetime string that specifies the date.
• weeks to add: An optional integer expression specifying the number of months to add to the date before calculating the first day of the week.

				  SELECT FDWEEK('02-08-2018');
				  -- Result: 2/4/2018

          SELECT FDWEEK('02-08-2018', 1)
          --Result: 02/11/2018
				  

FDMONTH(date)

• date: The datetime string that specifies the date.
• month to add: An optional integer expression specifying the number of months to add to the date before calculating the first day of the month.

				  SELECT FDMONTH('02-08-2018');
				  -- Result: 2/1/2018

          SELECT FDMONTH('02-08-2018', 1) 
          --Result: 03/01/2018
				  

FDQUARTER(date)

• date: The datetime string that specifies the date.
• quarters to add: An optional integer expression specifying the number of months to add to the date before calculating the first day of the quarter.

				  SELECT FDQUARTER('05-08-2018');
				  -- Result: 4/1/2018

          SELECT FDQUARTER('05-08-2018',1)
          --Result: 07/01/2018
				  

FILEMODIFIEDTIME(uri)

Returns the time stamp associated with the Date Modified of the relevant file.

• uri: An absolute path pointing to a file on the local file system.

				 SELECT FILEMODIFIEDTIME('C:/Documents/myfile.txt');
				 -- Result: 6/25/2019 10:06:58 AM
				 

FROM_DAYS(datevalue)

Returns a date derived from the number of days after 1582-10-15 (based upon the Gregorian calendar). This will be equivalent to the MYSQL FROM_DAYS function.

• datevalue: A integer value representing the number of days since 1582-10-15.

				SELECT FROM_DAYS(736000);
				-- Result: 2/6/2015
				

FROM_UNIXTIME(time, issecond)

Returns a representation of the unix_timestamp argument as a value in YYYY-MM-DD HH:MM:SS expressed in the current time zone.

• time: The time stamp value from epoch time. Milliseconds are accepted.
• issecond: Indicates the time stamp value is milliseconds to epoch time.

                      SELECT FROM_UNIXTIME(1540495231, 1);
                      -- Result: 2018-10-25 19:20:31

                      SELECT FROM_UNIXTIME(1540495357385, 0);
                      -- Result: 2018-10-25 19:22:37
                    

GETDATE()

Returns the current time stamp of the database system as a datetime value. This value is equal to CURRENT_TIMESTAMP and SYSDATETIME, and is always in the local timezone.

                  SELECT GETDATE();
                  -- Result: 2018-02-01 03:04:05
                

GETUTCDATE()

Returns the current time stamp of the database system formatted as a UTC datetime value. This value is equal to SYSUTCDATETIME.

In addition, GETUTCDATE can take an optional second parameter, a date and time that are converted to UTC.

                  SELECT GETUTCDATE();
                  -- For example, if the local timezone is Eastern European Time (GMT+2)
                  -- Result: 2018-02-01 05:04:05

                  SELECT GETUTCDATE('2020/08/31 13:56:00')
                  --Result: '2020-08-31 17:56:00'
                

HOUR(date)

Returns the hour component from the provided datetime.

• date: The datetime string that specifies the date.

				SELECT HOUR('02-02-2020 11:30:00');
				-- Result: 11
				

ISDATE(date, [date_format])

Returns 1 if the value is a valid date, time, or datetime value; otherwise, 0.

• date: The datetime string.
• date_format: The optional datetime format.

                      SELECT ISDATE('2018-02-01', 'yyyy-MM-dd');
                      -- Result: 1

                      SELECT ISDATE('Not a date');
                      -- Result: 0
                    

LAST_WEEK()

Returns a time stamp equivalent to exactly one week before the current date.

				SELECT LAST_WEEK();	//Assume the date is 3/17/2020	
			 -- Result: 3/10/2020 00:00:00
				

LAST_MONTH()

Returns a time stamp equivalent to exactly one month before the current date.

	
				SELECT LAST_MONTH(); //Assume the date is 3/17/2020
				-- Result: 2/17/2020 00:00:00
				

LAST_YEAR()

Returns a time stamp equivalent to exactly one year before the current date.

				SELECT LAST_YEAR();	//Assume the date is 3/17/2020	
				-- Result: 3/10/2019 00:00:00
				

LDWEEK(date)

Returns the last day of the provided week.

• date: The datetime string.
• weeks to add: An optional integer expression specifying the number of months to add to the date before calculating the last day of the week.

				SELECT LDWEEK('02-02-2020');
				-- Result: 2/8/2020
				

LDMONTH(date)

Returns the last day of the provided month.

• date: The datetime string.
• months to add: An optional integer expression specifying the number of months to add to the date before calculating the last day of the month.

				SELECT LDMONTH('02-02-2020');
				-- Result: 2/29/2020

        SELECT LDMONTH('02-08-2020', 1)
        --Result: 03/31/2020
				

LDQUARTER(date)

Returns the last day of the provided quarter.

• date: The datetime string.
• quarters to add: An optional integer expression specifying the number of months to add to the date before calculating the last day of the quarter.

				SELECT LDQUARTER('02-02-2020');
				-- Result: 3/31/2020

        SELECT LDQUARTER('02-02-2020',1)
        --Result: 06/30/2020
				

MAKEDATE(year, days)

Returns a date value from a year and a number of days.

• year: The year
• days: The number of days into the year. Value must be greater than 0.

          SELECT MAKEDATE(2020, 1);
          -- Result: 2020-01-01
        

MINUTE(date)

Returns the minute component from the provided datetime.

• date: The datetime string that specifies the date.

				SELECT MINUTE('02-02-2020 11:15:00');
				-- Result: 15
				

MONTH(date)

Returns the month component from the provided datetime.

• date: The datetime string that specifies the date.

				SELECT MONTH('02-02-2020');
				-- Result: 2
				

QUARTER(date)

Returns the quarter associated with the provided datetime.

• date: The datetime string that specifies the date.

				SELECT QUARTER('02-02-2020');
				-- Result: 1
				

SECOND(date)

Returns the second component from the provided datetime.

• date: The datetime string that specifies the date.

				SELECT SECOND('02-02-2020 11:15:23');
				-- Result: 23
				

SMALLDATETIMEFROMPARTS(integer_year, integer_month, integer_day, integer_hour, integer_minute)

Returns the datetime value for the specified date and time.

• year: The integer expression specifying the year.
• month: The integer expression specifying the month.
• day: The integer expression specifying the day.
• hour: The integer expression specifying the hour.
• minute: The integer expression specifying the minute.

                      SELECT SMALLDATETIMEFROMPARTS(2018, 2, 1, 1, 2);
                      -- Result: 2018-02-01 01:02:00
                    

STRTODATE(string,format)

Parses the provided string value and returns the corresponding datetime.

• string: The string value to be converted to datetime format.
• format: A format string which describes how to interpret the first string input. A few special formats are available as well, including UNIX, UNIXMILIS, TICKS, and FILETICKS.

				SELECT STRTODATE('03*04*2020','dd*MM*yyyy');
				-- Result: 4/3/2020
				

SYSDATETIME()

Returns the current time stamp as a datetime value of the database system. It is equal to GETDATE and CURRENT_TIMESTAMP, and is always in the local timezone.

                  SELECT SYSDATETIME();
                  -- Result: 2018-02-01 03:04:05
                

SYSUTCDATETIME()

Returns the current system date and time as a UTC datetime value. It is equal to GETUTCDATE.

                  SELECT SYSUTCDATETIME();
                  -- For example, if the local timezone is Eastern European Time (GMT+2)
                  -- Result: 2018-02-01 05:04:05
                

TIMEFROMPARTS(integer_hour, integer_minute, integer_seconds, integer_fractions, integer_precision)

Returns the time value for the specified time and with the specified precision.

• hour: The integer expression specifying the hour.
• minute: The integer expression specifying the minute.
• seconds: The integer expression specifying the seconds.
• fractions: The integer expression specifying the fractions of the second.
• precision : The integer expression specifying the precision of the fraction.

                      SELECT TIMEFROMPARTS(1, 2, 3, 456, 3);
                      -- Result: 01:02:03.456
                    

TO_DAYS(date)

Returns the number of days since 0000-00-01. This will only return a value for dates on or after 1582-10-15 (based upon the Gregorian calendar). This will be equivalent to the MYSQL TO_DAYS function.

• date: The datetime string that specifies the date.

				SELECT TO_DAYS('02-06-2015');
				-- Result: 736000
				

WEEK(date)

Returns the week (of the year) associated with the provided datetime.

• date: The datetime string that specifies the date.

				SELECT WEEK('02-17-2020 11:15:23');
				-- Result: 8
				

YEAR(date)

Returns the integer that specifies the year of the specified date.

• date: The datetime string.

                      SELECT YEAR('2018-02-01');
                      -- Result: 2018
                    

The following date literal functions can be used to filter date fields using relative intervals. Note that while the <, >, and = operators are supported for these functions, <= and >= are not.

L_TODAY()

The current day.

  SELECT * FROM MyTable WHERE MyDateField = L_TODAY()

L_YESTERDAY()

The previous day.

  SELECT * FROM MyTable WHERE MyDateField = L_YESTERDAY()

L_TOMORROW()

The following day.

  SELECT * FROM MyTable WHERE MyDateField = L_TOMORROW()

L_LAST_WEEK()

Every day in the preceding week.

  SELECT * FROM MyTable WHERE MyDateField = L_LAST_WEEK()

L_THIS_WEEK()

Every day in the current week.

  SELECT * FROM MyTable WHERE MyDateField = L_THIS_WEEK()

L_NEXT_WEEK()

Every day in the following week.

  SELECT * FROM MyTable WHERE MyDateField = L_NEXT_WEEK()

• L_LAST/L_THIS/L_NEXT MONTH
• L_LAST/L_THIS/L_NEXT QUARTER
• L_LAST/L_THIS/L_NEXT YEAR

L_LAST_N_DAYS(n)

The previous n days, excluding the current day.

  SELECT * FROM MyTable WHERE MyDateField = L_LAST_N_DAYS(3)

L_NEXT_N_DAYS(n)

The following n days, including the current day.

  SELECT * FROM MyTable WHERE MyDateField = L_NEXT_N_DAYS(3)

• L_LAST/L_NEXT_90_DAYS

L_LAST_N_WEEKS(n)

Every day in every week, starting n weeks before current week, and ending in the previous week.

  SELECT * FROM MyTable WHERE MyDateField = L_LAST_N_WEEKS(3)

L_NEXT_N_WEEKS(n)

Every day in every week, starting the following week, and ending n weeks in the future.

  SELECT * FROM MyTable WHERE MyDateField = L_NEXT_N_WEEKS(3)

• L_LAST/L_NEXT_N_MONTHS(n)
• L_LAST/L_NEXT_N_QUARTERS(n)
• L_LAST/L_NEXT_N_YEARS(n)

ABS(numeric_expr)

Returns the absolute (positive) value of the specified numeric expression.

• numeric_expr: A numeric expression.

ACOS(numeric_expr)

Returns the angle, in radians, whose cosine is the specified numeric expression; also called arccosine.

• numeric_expr: A numeric expression.

ASIN(numeric_expr)

Returns the angle, in radians, whose sine is the specified numeric expression. This is also called arcsine.

• numeric_expr: A numeric expression.

ATAN(numeric_expr)

Returns the angle, in radians, whose tangent is the specified numeric expression. This is also called arctangent.

• numeric_expr: A numeric expression.

CEILING(numeric_expr)

Returns the smallest integer value greater than, or equal to, the specified numeric expression.

• numeric_expr: A numeric expression.

COS(numeric_expr)

Returns the trigonometric cosine of the specified angle, in radians, in the specified expression.

• numeric_expr: A numeric expression.

COT(numeric_expr)

Returns the trigonometric cotangent of the specified angle, in radians, in the specified numeric expression.

• numeric_expr: A numeric expression.

DEGREES(numeric_expr)

Returns the corresponding angle in degrees for an angle specified in radians.

• numeric_expr: A numeric expression.

FLOOR(numeric_expr)

Returns the largest integer less than or equal to the specified numeric expression.

• numeric_expr: A numeric expression.

EXP(numeric_expr)

Returns the exponential value of the specified numeric expression.

• numeric_expr: A numeric expression.

LOG10(numeric_expr)

Returns the base-10 logarithm of the specified numeric expression.

• numeric_expr: A numeric expression.

RADIANS(numeric_expr)

Returns radians when a numeric expression, in degrees, is entered.

• numeric_expr: A numeric expression.

RAND()

Returns a randomly generated numeric value from [0,1).

ROUND(numeric_expr)

Returns a numeric value, rounded to the closest integer value.

• numeric_expr: A numeric expression.

SIGN(numeric_expr)

Returns the positive (+1), zero (0), or negative (-1) sign of the specified numeric expression.

• numeric_expr: A numeric expression.

SIN(numeric_expr)

Returns the trigonometric sine of the specified angle, in radians, in the specified expression.

• numeric_expr: A numeric expression.

SQRT(numeric_expr)

Returns the square root of the specified numeric value.

• numeric_expr: A numeric expression.

SQUARE(numeric_expr)

Returns the square of the specified numeric value.

• numeric_expr: A numeric expression.

TAN(numeric_expr)

Returns the tangent of the specified angle, in radians, in the specified expression.

• numeric_expr: A numeric expression.

TRUNC(numeric_expr)

Returns a numeric value, truncated to the closest integer value.

• numeric_expr: A numeric expression.

ATN2(y_expr, x_expr)

Returns the principal value of the arc tangent of y/x, expressed in radians.

• y_expr: The y numeric expression.
• x_expr: The x numeric expression.

LOG(numeric_expr [, base])

Returns the natural logarithm of the specified numeric expression.

• numeric_expr: A numeric expression.
• base: Optional numeric argument that sets the base for the logarithm.

PI()

Returns the constant value of PI.

POWER(numeric_expr, power_expr)

Returns the value of the specified expression to the specified power.

• numeric_expr: A numeric expression.
• power_expr: Is the power to which to raise numeric_expr.

IS_ARRAY(expr)

Returns a Boolean value indicating if the type of the specified expression is an array.

• expr: Any valid expression.

IS_BOOL(expr)

Returns a Boolean value indicating if the type of the specified expression is a Boolean.

• expr: Any valid expression.

IS_DEFINED(expr)

Returns a Boolean indicating if the property has been assigned a value.

• expr: Any valid expression.

IS_NULL(expr)

Returns a Boolean value indicating if the type of the specified expression is null.

• expr: Any valid expression.

IS_NUMBER(expr)

Returns a Boolean value indicating if the type of the specified expression is a number.

• expr: Any valid expression.

IS_OBJECT(expr)

Returns a Boolean value indicating if the type of the specified expression is a JSON object.

• expr: Any valid expression.

IS_PRIMITIVE(expr)

Returns a Boolean value indicating if the type of the specified expression is a primitive (string, Boolean, numeric, or null).

• expr: Any valid expression.

IS_STRING(expr)

Returns a Boolean value indicating if the type of the specified expression is a string.

• expr: Any valid expression.

CONCAT(str1, str2 [, str3] [, ...])

Returns a string that is the result of concatenating two or more string values.

• str1: The first string to concatenate.
• str2: The second string to concatenate.
• str3: The third string to concatenate.

CONTAINS(str1, str2)

Returns a Boolean indicating whether the first string expression contains the second.

• str1: The string to search in.
• str2: The string to search for.

ENDSWITH(str1, str2)

Returns a Boolean indicating whether the first string expression ends with the second.

• str1: The string to search in.
• str2: The string to search for.

INDEX_OF(str1, str2)

Returns the starting position of the first occurrence of the second string expression within the first specified string expression, or -1 if the string is not found.

• str1: The string to search in.
• str2: The string to search for.

LEFT(str, num_expr)

Returns the left part of a string with the specified number of characters.

• str: A valid string expression.
• num_expr: The number of characters to return.

LENGTH(str)

Returns the number of characters of the specified string expression.

• str: Any valid string expression.

LOWER(str)

Returns a string expression after converting uppercase character data to lowercase.

• str: Any valid string expression.

LTRIM(str)

Returns a string expression after it removes leading blanks.

• str: Any valid string expression.

REPLACE(original_value, from_value, to_value)

Replaces all occurrences of a specified string value with another string value.

• original_value: The string to search in.
• from_value: The string to search for.
• to_value: The string to replace instances of from_value.

REPLICATE(str, repeat_num)

Repeats a string value a specified number of times.

• str: The string expression to repeat.
• repeat_num: The number of times to repeat the str expression.

REVERSE(str)

Returns the reverse order of a string value.

• str: Any valid string expression.

RIGHT(str, num_expr)

Returns the right part of a string with the specified number of characters.

• str: Any valid string expression.
• num_expr: The starting index.

RTRIM(str)

Returns a string expression after it removes trailing blanks.

• str: Any valid string expression.

STARTSWITH(str1, str2)

Returns a Boolean indicating whether the first string expression starts with the second.

• str1: The string to search in.
• str2: The string to search for.

SUBSTRING(str, start_index, length)

Returns part of a string expression starting at the specified character zero-based position and continues to the specified length, or to the end of the string.

• str: Any valid string expression.
• start_index: The starting index.
• length: The length of the string to return.

TOSTRING(expr)

Returns a string representation of scalar expression.

• expr: Any valid expression.

TRIM(str)

Returns a string expression after it removes leading and trailing blanks.

• str: Any valid string expression.

UPPER(str)

Returns a string expression after converting lowercase character data to uppercase.

• str: Any valid string expression.

ARRAY_CONCAT(array_exp1, array_exp2 [, array_exp3])

Returns an array that is the result of concatenating two or more array values.

• array_exp1: Any valid array expression.
• array_exp2: Any valid array expression.
• array_exp3: Any valid array expression.

ARRAY_CONTAINS(array_exp, expr [, bool_expr])

Returns a Boolean indicating whether the array contains the specified value. You can check for a partial or full match of an object by using a boolean expression within the command.

• array_exp1: Any array expression.
• expr: The expression to search for.
• bool_expr: If it's set to 'true'and if the specified search value is an object, the command checks for a partial match (the search object is a subset of one of the objects). If it's set to 'false', the command checks for a full match of all objects within the array. The default value if not specified is false.

ARRAY_LENGTH(array_exp)

Returns the number of elements of the specified array expression.

• array_exp: Any valid array expression.

ARRAY_SLICE(array_exp, start_index, max_size)

Returns part of an array expression.

• array_exp: Any valid array expression.
• start_index: Zero-based numeric index at which to begin the array. Negative values may be used to specify the starting index relative to the last element of the array i.e. -1 references the last element in the array.
• max_size: Maximum number of elements in the resulting array.

ST_DISTANCE(spatial_expr1, spatial_expr2)

Returns the distance between the two GeoJSON Point, Polygon, or LineString expressions.

• spatial_expr1: Is any valid GeoJSON Point, Polygon, or LineString object expression.
• spatial_expr2: Is any valid GeoJSON Point, Polygon, or LineString object expression.

ST_WITHIN(spatial_expr1, spatial_expr2)

Returns a Boolean expression indicating whether the GeoJSON object (Point, Polygon, or LineString) specified in the first argument is within the GeoJSON (Point, Polygon, or LineString) in the second argument.

• spatial_expr1: Is any valid GeoJSON Point, Polygon, or LineString object expression.
• spatial_expr2: Is any valid GeoJSON Point, Polygon, or LineString object expression.

ST_INTERSECTS(1, 2)

Returns a Boolean expression indicating whether the GeoJSON object (Point, Polygon, or LineString) specified in the first argument intersects the GeoJSON (Point, Polygon, or LineString) in the second argument.

• spatial_expr1: Is any valid GeoJSON Point, Polygon, or LineString object expression.
• spatial_expr2: Is any valid GeoJSON Point, Polygon, or LineString object expression.

ST_ISVALID(spatial_expr)

Returns a Boolean value indicating whether the specified GeoJSON Point, Polygon, or LineString expression is valid.

• spatial_expr: Is any valid GeoJSON Point, Polygon, or LineString object expression.

ST_ISVALIDDETAILED(spatial_expr)

Returns a JSON value containing a Boolean value if the specified GeoJSON Point, Polygon, or LineString expression is valid, and if invalid, additionally the reason as a string value.

• spatial_expr: Is any valid GeoJSON Point, Polygon, or LineString object expression.

ABS(numeric_expr)

Returns the absolute (positive) value of the specified numeric expression.

• numeric_expr: A numeric expression.

ACOS(numeric_expr)

Returns the angle, in radians, whose cosine is the specified numeric expression; also called arccosine.

• numeric_expr: A numeric expression.

ASIN(numeric_expr)

Returns the angle, in radians, whose sine is the specified numeric expression. This is also called arcsine.

• numeric_expr: A numeric expression.

ATAN(numeric_expr)

Returns the angle, in radians, whose tangent is the specified numeric expression. This is also called arctangent.

• numeric_expr: A numeric expression.

CEILING(numeric_expr)

Returns the smallest integer value greater than, or equal to, the specified numeric expression.

• numeric_expr: A numeric expression.

COS(numeric_expr)

Returns the trigonometric cosine of the specified angle, in radians, in the specified expression.

• numeric_expr: A numeric expression.

COT(numeric_expr)

Returns the trigonometric cotangent of the specified angle, in radians, in the specified numeric expression.

• numeric_expr: A numeric expression.

DEGREES(numeric_expr)

Returns the corresponding angle in degrees for an angle specified in radians.

• numeric_expr: A numeric expression.

FLOOR(numeric_expr)

Returns the largest integer less than or equal to the specified numeric expression.

• numeric_expr: A numeric expression.

EXP(numeric_expr)

Returns the exponential value of the specified numeric expression.

• numeric_expr: A numeric expression.

LOG10(numeric_expr)

Returns the base-10 logarithm of the specified numeric expression.

• numeric_expr: A numeric expression.

RADIANS(numeric_expr)

Returns radians when a numeric expression, in degrees, is entered.

• numeric_expr: A numeric expression.

RAND()

Returns a randomly generated numeric value from [0,1).

ROUND(numeric_expr)

Returns a numeric value, rounded to the closest integer value.

• numeric_expr: A numeric expression.

SIGN(numeric_expr)

Returns the positive (+1), zero (0), or negative (-1) sign of the specified numeric expression.

• numeric_expr: A numeric expression.

SIN(numeric_expr)

Returns the trigonometric sine of the specified angle, in radians, in the specified expression.

• numeric_expr: A numeric expression.

SQRT(numeric_expr)

Returns the square root of the specified numeric value.

• numeric_expr: A numeric expression.

SQUARE(numeric_expr)

Returns the square of the specified numeric value.

• numeric_expr: A numeric expression.

TAN(numeric_expr)

Returns the tangent of the specified angle, in radians, in the specified expression.

• numeric_expr: A numeric expression.

TRUNC(numeric_expr)

Returns a numeric value, truncated to the closest integer value.

• numeric_expr: A numeric expression.

ATN2(y_expr, x_expr)

Returns the principal value of the arc tangent of y/x, expressed in radians.

• y_expr: The y numeric expression.
• x_expr: The x numeric expression.

LOG(numeric_expr [, base])

Returns the natural logarithm of the specified numeric expression.

• numeric_expr: A numeric expression.
• base: Optional numeric argument that sets the base for the logarithm.

PI()

Returns the constant value of PI.

POWER(numeric_expr, power_expr)

Returns the value of the specified expression to the specified power.

• numeric_expr: A numeric expression.
• power_expr: Is the power to which to raise numeric_expr.

IS_ARRAY(expr)

Returns a Boolean value indicating if the type of the specified expression is an array.

• expr: Any valid expression.

IS_BOOL(expr)

Returns a Boolean value indicating if the type of the specified expression is a Boolean.

• expr: Any valid expression.

IS_DEFINED(expr)

Returns a Boolean indicating if the property has been assigned a value.

• expr: Any valid expression.

IS_NULL(expr)

Returns a Boolean value indicating if the type of the specified expression is null.

• expr: Any valid expression.

IS_NUMBER(expr)

Returns a Boolean value indicating if the type of the specified expression is a number.

• expr: Any valid expression.

IS_OBJECT(expr)

Returns a Boolean value indicating if the type of the specified expression is a JSON object.

• expr: Any valid expression.

IS_PRIMITIVE(expr)

Returns a Boolean value indicating if the type of the specified expression is a primitive (string, Boolean, numeric, or null).

• expr: Any valid expression.

IS_STRING(expr)

Returns a Boolean value indicating if the type of the specified expression is a string.

• expr: Any valid expression.

CONCAT(str1, str2 [, str3] [, ...])

Returns a string that is the result of concatenating two or more string values.

• str1: The first string to concatenate.
• str2: The second string to concatenate.
• str3: The third string to concatenate.

CONTAINS(str1, str2)

Returns a Boolean indicating whether the first string expression contains the second.

• str1: The string to search in.
• str2: The string to search for.

ENDSWITH(str1, str2)

Returns a Boolean indicating whether the first string expression ends with the second.

• str1: The string to search in.
• str2: The string to search for.

INDEX_OF(str1, str2)

Returns the starting position of the first occurrence of the second string expression within the first specified string expression, or -1 if the string is not found.

• str1: The string to search in.
• str2: The string to search for.

LEFT(str, num_expr)

Returns the left part of a string with the specified number of characters.

• str: A valid string expression.
• num_expr: The number of characters to return.

LENGTH(str)

Returns the number of characters of the specified string expression.

• str: Any valid string expression.

LOWER(str)

Returns a string expression after converting uppercase character data to lowercase.

• str: Any valid string expression.

LTRIM(str)

Returns a string expression after it removes leading blanks.

• str: Any valid string expression.

REPLACE(original_value, from_value, to_value)

Replaces all occurrences of a specified string value with another string value.

• original_value: The string to search in.
• from_value: The string to search for.
• to_value: The string to replace instances of from_value.

REPLICATE(str, repeat_num)

Repeats a string value a specified number of times.

• str: The string expression to repeat.
• repeat_num: The number of times to repeat the str expression.

REVERSE(str)

Returns the reverse order of a string value.

• str: Any valid string expression.

RIGHT(str, num_expr)

Returns the right part of a string with the specified number of characters.

• str: Any valid string expression.
• num_expr: The starting index.

RTRIM(str)

Returns a string expression after it removes trailing blanks.

• str: Any valid string expression.

STARTSWITH(str1, str2)

Returns a Boolean indicating whether the first string expression starts with the second.

• str1: The string to search in.
• str2: The string to search for.

SUBSTRING(str, start_index, length)

Returns part of a string expression starting at the specified character zero-based position and continues to the specified length, or to the end of the string.

• str: Any valid string expression.
• start_index: The starting index.
• length: The length of the string to return.

TOSTRING(expr)

Returns a string representation of scalar expression.

• expr: Any valid expression.

TRIM(str)

Returns a string expression after it removes leading and trailing blanks.

• str: Any valid string expression.

UPPER(str)

Returns a string expression after converting lowercase character data to uppercase.

• str: Any valid string expression.

ARRAY_CONCAT(array_exp1, array_exp2 [, array_exp3])

Returns an array that is the result of concatenating two or more array values.

• array_exp1: Any valid array expression.
• array_exp2: Any valid array expression.
• array_exp3: Any valid array expression.

ARRAY_CONTAINS(array_exp, expr [, bool_expr])

Returns a Boolean indicating whether the array contains the specified value. You can check for a partial or full match of an object by using a boolean expression within the command.

• array_exp1: Any array expression.
• expr: The expression to search for.
• bool_expr: If it's set to 'true'and if the specified search value is an object, the command checks for a partial match (the search object is a subset of one of the objects). If it's set to 'false', the command checks for a full match of all objects within the array. The default value if not specified is false.

ARRAY_LENGTH(array_exp)

Returns the number of elements of the specified array expression.

• array_exp: Any valid array expression.

ARRAY_SLICE(array_exp, start_index, max_size)

Returns part of an array expression.

• array_exp: Any valid array expression.
• start_index: Zero-based numeric index at which to begin the array. Negative values may be used to specify the starting index relative to the last element of the array i.e. -1 references the last element in the array.
• max_size: Maximum number of elements in the resulting array.

ST_DISTANCE(spatial_expr1, spatial_expr2)

Returns the distance between the two GeoJSON Point, Polygon, or LineString expressions.

• spatial_expr1: Is any valid GeoJSON Point, Polygon, or LineString object expression.
• spatial_expr2: Is any valid GeoJSON Point, Polygon, or LineString object expression.

ST_WITHIN(spatial_expr1, spatial_expr2)

Returns a Boolean expression indicating whether the GeoJSON object (Point, Polygon, or LineString) specified in the first argument is within the GeoJSON (Point, Polygon, or LineString) in the second argument.

• spatial_expr1: Is any valid GeoJSON Point, Polygon, or LineString object expression.
• spatial_expr2: Is any valid GeoJSON Point, Polygon, or LineString object expression.

ST_INTERSECTS(1, 2)

Returns a Boolean expression indicating whether the GeoJSON object (Point, Polygon, or LineString) specified in the first argument intersects the GeoJSON (Point, Polygon, or LineString) in the second argument.

• spatial_expr1: Is any valid GeoJSON Point, Polygon, or LineString object expression.
• spatial_expr2: Is any valid GeoJSON Point, Polygon, or LineString object expression.

ST_ISVALID(spatial_expr)

Returns a Boolean value indicating whether the specified GeoJSON Point, Polygon, or LineString expression is valid.

• spatial_expr: Is any valid GeoJSON Point, Polygon, or LineString object expression.

ST_ISVALIDDETAILED(spatial_expr)

Returns a JSON value containing a Boolean value if the specified GeoJSON Point, Polygon, or LineString expression is valid, and if invalid, additionally the reason as a string value.

• spatial_expr: Is any valid GeoJSON Point, Polygon, or LineString object expression.

A SELECT statement can consist of the following basic clauses.

• SELECT
• INTO
• FROM
• JOIN
• WHERE
• GROUP BY
• HAVING
• UNION
• ORDER BY
• LIMIT

SELECT Syntax

The following syntax diagram outlines the syntax supported by the SQL engine of the connector:

SELECT {
  [ TOP <numeric_literal> | DISTINCT ]
  { 
    * 
    | { 
        <expression> [ [ AS ] <column_reference> ] 
        | { <table_name> | <correlation_name> } .* 
      } [ , ... ] 
  }
  { 
    FROM <table_reference> [ [ AS ] <identifier> ] 
  } [ , ... ]
  [ [  
      INNER | { { LEFT | RIGHT | FULL } [ OUTER ] } 
    ] JOIN <table_reference> [ ON <search_condition> ] [ [ AS ] <identifier> ] 
  ] [ ... ] 
  [ WHERE <search_condition> ]
  [ GROUP BY <column_reference> [ , ... ]
  [ HAVING <search_condition> ]
  [ UNION [ ALL ] <select_statement> ]
  [ 
    ORDER BY 
    <column_reference> [ ASC | DESC ] [ NULLS FIRST | NULLS LAST ]
  ]
  [ 
    LIMIT <expression>
    [ 
      { OFFSET | , }
      <expression> 
    ]
  ] 
} | SCOPE_IDENTITY() 

<expression> ::=
  | <column_reference>
  | @ <parameter> 
  | ?
  | COUNT( * | { [ DISTINCT ] <expression> } )
  | { AVG | MAX | MIN | SUM | COUNT } ( <expression> ) 
  | NULLIF ( <expression> , <expression> ) 
  | COALESCE ( <expression> , ... ) 
  | CASE <expression>
      WHEN { <expression> | <search_condition> } THEN { <expression> | NULL } [ ... ]
    [ ELSE { <expression> | NULL } ]
    END 
  | {RANK() | DENSE_RANK()} OVER ([PARTITION BY <column_reference>] {ORDER BY <column_reference>})
  | <literal>
  | <sql_function> 

<search_condition> ::= 
  {
    <expression> { = | > | < | >= | <= | <> | != | LIKE | NOT LIKE | IN | NOT IN | IS NULL | IS NOT NULL | AND | OR | CONTAINS | BETWEEN | IS DISTINCT FROM | IS NOT DISTINCT FROM } [ <expression> ]
  } [ { AND | OR } ... ] 

Examples

1. Return all columns:
   

   SELECT * FROM [CData].[Entities].Customers

2. Rename a column:
   

   SELECT [CompanyName] AS MY_CompanyName FROM [CData].[Entities].Customers

3. Cast a column's data as a different data type:
   

   SELECT CAST(Balance AS VARCHAR) AS Str_Balance FROM [CData].[Entities].Customers

4. Search data:
   

   SELECT * FROM [CData].[Entities].Customers WHERE Country = 'US'

5. Return the number of items matching the query criteria:
   

   SELECT COUNT(*) AS MyCount FROM [CData].[Entities].Customers 

6. Return the number of unique items matching the query criteria:
   

   SELECT COUNT(DISTINCT CompanyName) FROM [CData].[Entities].Customers 

7. Return the unique items matching the query criteria:
   

   SELECT DISTINCT CompanyName FROM [CData].[Entities].Customers 

8. Sort a result set in ascending order:
   

   SELECT City, CompanyName FROM [CData].[Entities].Customers  ORDER BY CompanyName ASC

9. Restrict a result set to the specified number of rows:
   

   SELECT City, CompanyName FROM [CData].[Entities].Customers LIMIT 10 

10. Parameterize a query to pass in inputs at execution time. This enables you to create prepared statements and mitigate SQL injection attacks.
    

    SELECT * FROM [CData].[Entities].Customers WHERE Country = @param

Pseudo Columns

Some input-only fields are available in SELECT statements. These fields, called pseudo columns, do not appear as regular columns in the results, yet may be specified as part of the WHERE clause. You can use pseudo columns to access additional features from Azure Cosmos DB.

    SELECT * FROM [CData].[Entities].Customers WHERE MyPseudocolumn = 'MyValue'
    

Aggregate Functions

For SELECT examples using aggregate functions, see Aggregate Functions.

JOIN Queries

See JOIN Queries for SELECT query examples using JOINs.

Date Literal Functions

Date Literal Functions contains SELECT examples with date literal functions.

Projection Functions

See Projection Functions for SELECT examples with projection functions.

Predicate Functions

For SELECT examples using predicate functions, see Predicate Functions.

Window Functions

See Window Functions for SELECT examples containing window functions.

Table-Valued Functions

See Table-Valued Functions for SELECT examples with table-valued functions.

COUNT

Returns the number of rows matching the query criteria.

SELECT COUNT(*) FROM [CData].[Entities].Customers WHERE Country = 'US'

COUNT(DISTINCT)

Returns the number of distinct, non-null field values matching the query criteria.

SELECT COUNT(DISTINCT City) AS DistinctValues FROM [CData].[Entities].Customers WHERE Country = 'US'

AVG

Returns the average of the column values.

SELECT CompanyName, AVG(Balance) FROM [CData].[Entities].Customers WHERE Country = 'US'  GROUP BY CompanyName

MIN

Returns the minimum column value.

SELECT MIN(Balance), CompanyName FROM [CData].[Entities].Customers WHERE Country = 'US' GROUP BY CompanyName

MAX

Returns the maximum column value.

SELECT CompanyName, MAX(Balance) FROM [CData].[Entities].Customers WHERE Country = 'US' GROUP BY CompanyName

SUM

Returns the total sum of the column values.

SELECT SUM(Balance) FROM [CData].[Entities].Customers WHERE Country = 'US'

The CData Python Connector for Azure Cosmos DB supports joins of a nested array with its parent document and joins of multiple collections.

Joining Nested Structures

The connector expects the left part of the join is the array document you want to flatten vertically. Set SupportEnhancedSQL to false to join nested Azure Cosmos DB documents. This type of query is supported through the Azure Cosmos DB API.

For example, consider the following query from Azure Cosmos DB's restaurants collection:

SELECT [restaurants].[restaurant_id], [restaurants].name, [restaurants.grades].* 
FROM [restaurants.grades]
JOIN [restaurants] 
WHERE [restaurants].name = 'Morris Park Bake Shop'

Joining Multiple Collections

You can join multiple collections just like you would join tables in a relational database. Set SupportEnhancedSQL to True to execute these types of joins. The following examples use the restaurants and zips collections available in the Azure Cosmos DB documentation.

The query below returns the restaurant records that exist, if any, for each ZIP code:

SELECT z.city, r.name, r.borough, r.cuisine, r.[address.zipcode]
FROM zips z
LEFT JOIN restaurants r
ON r.[address.zipcode] = z._id

The query below returns records from both tables that match the join condition:

SELECT z.city, r.name, r.borough, r.cuisine, r.[address.zipcode]
FROM restaurants r
INNER JOIN zips z
ON r.[address.zipcode] = z._id

Window functions allow you to create computed fields from a group of rows (a window) that return a result for each row, as opposed to one computed result for a set of rows, as is the case with aggregate functions. The connector supports the following window function syntax.

Note: Window function support is an experimental feature of the connector. This functionality extends beyond the connector's core scope of being SQL-92 compliant. As such, performance with window functions may not be optimal.

Window Function Clauses

OVER

The OVER clause defines the window over which window functions are performed.

SELECT A, B, <window function> OVER (<window frame>) FROM TableName

The <window function> refers to any supported window function clause, and the <window frame> refers to one or more clauses that specify the logic by which the window is defined.

PARTITION BY

The PARTITION BY clause subdivides a window into sub-windows called partitions. For each unique value in the column specified in the PARTITION BY clause, every record with that value collectively forms an individual partition.

SELECT A, B, <window function> OVER (PARTITION BY A ORDER BY B) From [CData].[Entities].Customers

The <window function> refers to any supported window function clause.

Window Functions

The connector supports math, ranking, and analytic window functions.

Math

These window functions perform mathematical operations on the records within the window.

COUNT()

Calculates the number of records in each partition. The calculated column is of the data type "int".

In each partition, every record will display the total number of records in that partition.

SELECT Name, Role, Earnings, COUNT() OVER (PARTITION BY Role) FROM Employees

COUNT_BIG()

Calculates the number of records in each partition. The calculated column is of the data type "bigint".

In each partition, every record will display the total number of records in that partition.

SELECT Name, Role, Earnings, COUNT_BIG() OVER (PARTITION BY Role) FROM Employees

MIN(numeric_column)

Calculates the minimum value of a numerical column per partition.

In each partition, every record will display the minimum value of numeric_column across the records in that partition.

SELECT Name, Role, Earnings, MIN(Earnings) OVER (PARTITION BY Role) FROM Employees

MAX(numeric_column)

Calculates the maximum value of a numerical column per partition.

In each partition, every record will display the maximum value of numeric_column across the records in that partition.

SELECT Name, Role, Earnings, MAX(Earnings) OVER (PARTITION BY Role) FROM Employees

SUM(numeric_column)

Calculates the sum of a numerical column per partition.

In each partition, every record will display the sum of numeric_column across the records in that partition.

SELECT Name, Role, Earnings, SUM(Earnings) OVER (PARTITION BY Role) FROM Employees

AVG(numeric_column)

Calculates the average value of a numerical column per partition.

In each partition, every record will display the average value of numeric_column across the records in that partition.

SELECT Name, Role, Earnings, AVG(Earnings) OVER (PARTITION BY Role) FROM Employees

MEDIAN(numeric_column)

Calculates the median value of a numerical column per partition.

In each partition, every record will display the median value of numeric_column across the records in that partition.

SELECT Name, Role, Earnings, MEDIAN(Earnings) OVER (PARTITION BY Role) FROM Employees

STDEV(numeric_column)

Calculates the standard deviation of a numerical column per partition.

In each partition, every record will display the standard deviation of numeric_column across the records in that partition.

SELECT Name, Role, Earnings, STDEV(Earnings) OVER (PARTITION BY Role) FROM Employees

STDEVP(numeric_column)

Calculates the population standard deviation of a numerical column per partition.

In each partition, every record will display the population standard deviation of numeric_column across the records in that partition.

SELECT Name, Role, Earnings, STDEVP(Earnings) OVER (PARTITION BY Role) FROM Employees

VAR(numeric_column)

Calculates the statistical standard variance of a numerical column per partition.

In each partition, every record will display the statistical standard variance of numeric_column across the records in that partition.

SELECT Name, Role, Earnings, VAR(Earnings) OVER (PARTITION BY Role) FROM Employees

VARP(numeric_column)

Calculates the variance population of a numerical column per partition.

In each partition, every record will display the variance population of numeric_column across the records in that partition.

SELECT Name, Role, Earnings, VARP(Earnings) OVER (PARTITION BY Role) FROM Employees

Ranking

These window functions rank records that fall within the window and its partitions.

RANK()

Assigns a rank number to each record in a window based on the value of the column specified in the required ORDER BY clause.

If two or more records have an equal value in the in ranked column, they all receive the same rank number and the rank count increments internally, skipping ahead one rank number for each record with a duplicate value in the ORDER BY column.

SELECT City, CompanyName, RANK() OVER (ORDER BY CompanyName) AS Rank FROM [CData].[Entities].Customers

If you add a PARTITION BY clause, a separate set of ranks is calculated for each partition.

SELECT City, CompanyName, RANK() OVER (PARTITION BY City ORDER BY CompanyName) AS Rank FROM [CData].[Entities].Customers

DENSE_RANK()

Operates like the RANK() function, but it doesn't increment the internal rank counter for each record with a duplicate value in the ranked column.

This means that, while records with identical values in the ORDER BY column still share a rank number, the function never skips a rank number.

SELECT City, CompanyName, DENSE_RANK() OVER (PARTITION BY City ORDER BY CompanyName) AS Rank FROM [CData].[Entities].Customers

If you add a PARTITION BY clause, a separate set of ranks is calculated for each partition.

SELECT City, CompanyName, DENSE_RANK() OVER (PARTITION BY City ORDER BY CompanyName) AS Rank FROM [CData].[Entities].Customers

ROW_NUMBER()

Calculates a row number for each record. An ORDER BY clause in the OVER clause is required.

SELECT Name, Role, Earnings, ROW_NUMBER() OVER (ORDER BY Role) FROM Employees

SELECT Name, Role, Earnings, ROW_NUMBER() OVER (PARTITION BY Role ORDER BY Earnings) FROM Employees

NTILE()

Distributes rows of an ordered partition into a specified number of approximately equal groups, or buckets. It assigns each group a bucket number starting from one. For each row in a group, the NTILE() function assigns a bucket number representing the group to which the row belongs.

The syntax of NTILE() is:

NTILE(buckets) OVER (
    [PARTITION BY partition_expression, ... ]
    ORDER BY sort_expression [ASC | DESC], ...
)

• buckets: The number of buckets into which the rows are divided. The buckets can be an expression or subquery that evaluates to a positive integer. It cannot be a window function.
• PARTITION BY: distributes rows of a result set into partitions to which the NTILE() function is applied.
• ORDER BY is clause that specifies the logical order of rows in each partition to which the NTILE() is applied.

If the number of rows is not divisible by the buckets, the NTILE() function returns groups of two sizes with the difference by one. The larger groups always precede the smaller group in the order set by ORDER BY in the OVER() clause.

If the total of rows is divisible by the number of buckets, the function divides the rows evenly among buckets. The following statement creates a new table named ntile_demo that stores 10 integers:

CREATE TABLE sales.ntile_demo (
	v INT NOT NULL
);
	
INSERT INTO sales.ntile_demo(v) 
VALUES(1),(2),(3),(4),(5),(6),(7),(8),(9),(10);	
	
SELECT * FROM sales.ntile_demo;

SELECT 
	v, 
	NTILE (3) OVER (
		ORDER BY v
	) buckets
FROM 
	sales.ntile_demo;

Analytical

These window functions perform analytical operations on the records within the window.

PERCENT_RANK()

The syntax of PERCENT_RANK() is:

PERCENT_RANK() OVER (
      [PARTITION BY partition_expression, ... ]
      ORDER BY sort_expression [ASC | DESC], ...
  )
  

• PARTITION BY: By default, SQL Server treats the whole data set as a single set. You can specify the PARTITION BY clause to divide data into multiple sets. The Percent_Rank function performs the analytical calculations on each set. This parameter is optional.
• ORDER BY: Sorts the data in either ascending or descending order. This parameter is required.

Table-valued functions are functions that return a table (rowset).

Note: Table-valued function support is an experimental feature of the connector. This functionality extends beyond the connector's core scope of being SQL-92 compliant. As such, performance with these functions may not be optimal.

Table-Valued Function Clauses

CROSS APPLY

The CROSS APPLY operator is used to perform a subquery on each row of a table or resultset produced by a preceding table expression.

<table_expression_1> CROSS APPLY <table_expression_2>

The second table expression can reference results from the first table expression to create derived columns or an altered recordset via a table-valued function.

Each resulting record is an instance of the record it's splitting, with all the same column values, except for the column(s) containing the value split by the function.

WITH

SELECT A.ColumnName, X.DerivedColumnName FROM TableName A CROSS APPLY <table-valued function> WITH (DerivedColumnName varchar(255)) AS X

Table-Valued Functions

STRING_SPLIT(input_text,delimiter)

Takes each record in the recordset of the preceding table expression, splits the column containing delimiters (input_text) into substrings separated by the delimiter, and returns one record per substring.

• input_text: A column whose value you want to parse.
• delimiter: The character used to split the value of the column specified in input_text.

Suppose there is a column called "SplitColumn" with the following content:

One-Two-Three

SELECT A.ID, X.Value FROM [TableWithDelimitedStringField] A CROSS APPLY STRING_SPLIT(A.SplitColumn,'-') WITH (Value VARCHAR(255)) AS X

-- Results:
-----------
|ID|Value|
|1|One|
|1|Two|
|1|Three|

JSONTABLE(json_content,[jsonpath])

For each record in the recordset of the preceding table expression, returns one record for each instance of a key in a JSON array (json_content) that matches the key(s) specified in the WITH clause, at the scope specified by the "jsonpath" input.

• json_content: A JSON "table" (array of objects). The contents can nest, but this must be a single JSON array, not any other JSON structure, at the root level.
  • The values of every instance of the key(s) provided in the WITH clause are retrievable only for substructures which are immediate children of the root-level JSON array.
• jsonpath: An optional JSONPath query defining the scope, within the json_content array, that you want to pull content from. The JSON key(s) identified in the WITH clause must exist at the scope defined in this parameter. This defaults to the JSON root ($).

Consider a sample table with a single record, including an ID column and column with JSON content called "JSONColumn" with the following content:

[
	{
		"name": "Samuel",
		"email": "sam@gmail.com",
		"extrainfo": {
			"city": "Seattle"
		}
	},
	{
		"name": "Katherine",
		"email": "kat@gmail.com",
	},
	{
		"name": "George",
		"email": "george23@gmail.com",
	},
	{
		"name": "Carlos",
		"email": "carlos32@gmail.com",
	}
]

To extract all values for a certain key, specify the scope in the JSONTABLE function and provide the desired key(s) in the WITH clause.

SELECT A.ID, X.name FROM [TableWithJSONField] A CROSS APPLY JSONTABLE(A.JSONColumn) WITH (name VARCHAR(255)) AS X

-- Results: 
|ID|name|
---------
|1 |Samuel|
|1 |Katherine|
|1 |George|
|1 |Carlos|

XMLTABLE(xml_content,[xpath,child_type])

For each record in the resultset of the preceding table expression, returns one record for each of the elements and/or attributes in an XML structure (xml_content) that match the tag name(s) and/or attribute name(s) specified in the WITH clause, at the scope specified in the "xpath" input.

• xml_content: A column containing an XML structure.
• xpath: An optional XPath that specifies the scope within the XML structure at which the connector extracts content matching the tag/attribute name(s) specified in the WITH clause.
  • When extracting the content of sub-elements, the connector can retrieve all content from tags at the root level, (depth 0) immediate children of the root (depth 1), and children of those children (depth 2).
  • When extracting element attribute content, the connector can retrieve all content from tags containing the specified attribute at the root level (depth 0) and from immediate children of root-level elements (depth 1).
• child_type: An optional parameter that specifies the part(s) of the parent element (specified in the xpath input) that the column(s) provided in the WITH clause are checked against to identify content.
  • You can supply the following values:
    • 0: The column(s) in the WITH clause are checked for matches against the parent element's attribute names and sub-element tag names.
    • 1: The column(s) in the WITH clause are checked for matches against the parent element's attribute names.
    • 2: The column(s) in the WITH clause are checked for matches against the parent element's sub-element tag names.
  • When not supplied, this defaults to 0.

Extracting Sub-Element Values

<shoppingList>
    <item>
        <name>Apples</name>
        <quantity>3</quantity>
        <unit>Kg</unit>
    </item>
    <item>
        <name>Bread</name>
        <quantity>2</quantity>
        <unit>Loaf</unit>
		<extrainfo>
			<Type>Whole-Grain</Type>
		</extrainfo>
    </item>
    <item>
        <name>Milk</name>
        <quantity>1</quantity>
        <unit>Carton</unit>
    </item>
    <item>
        <name>Eggs</name>
        <quantity>12</quantity>
        <unit></unit>
    </item>
</shoppingList>

To extract sub-element content, specify the scope in the XMLTABLE function and provide the desired element name(s) in the WITH clause. Note that this will not work if the XMLTABLE function's child_type input is set to 1.

SELECT A.ID, X.name FROM [TableWithXMLField] A CROSS APPLY XMLTABLE(A.XMLContent,'//*/item') WITH (name VARCHAR(255)) AS X

-- Results: 
|ID|name|
---------
|1|Apples|
|1|Bread|
|1|Milk|
|1|Eggs|

Extracting Values Using Element Tag Attributes

Suppose you have this sample table with a single record, including an ID column and a column with XML content called "XMLContent" with the following content:

<restaurant>
  <dish type="appetizer">
    <name lang="en">Caprese Salad</name>
    <chef>Chef Giovanni</chef>
    <price currency="USD">9.99</price>
  </dish>
  <dish type="main-course">
    <name lang="fr">Boeuf Bourguignon</name>
    <chef>Chef Marie</chef>
    <price currency="EUR">19.99</price>
  </dish>
  <dish type="dessert">
    <name lang="es">Tres Leches Cake</name>
    <chef>Chef Alejandro</chef>
    <price currency="MXN">89.99</price>
  </dish>
</restaurant>

To extract attribute content, specify the scope in the XMLTABLE function and provide the desired attribute name(s) in the WITH clause. Note that this will not work if the XMLTABLE function's child_type input is set to 2.

SELECT A.ID, X.type FROM [TableWithXMLField] A CROSS APPLY XMLTABLE(A.XMLContent,'//*/dish') WITH (type VARCHAR(255)) AS X

-- Results: 
|ID|type|
---------
|1|appetizer|
|1|main-course|
|1|dessert|

CSVTABLE(csv_content,[delimiter])

For each record in the resultset of the preceding table expression, reads from a column that contains a CSV table (csv_content) and for each record in that CSV table, returns one record containing the value of the CSV column(s) specified in the WITH clause.

• csv_content: A column containing a CSV table.
• delimiter: An optional custom delimiter (instead of a comma) which splits the CSV content contained in the csv_content input.

Consider a sample table with a single record, including an ID column and a column containing CSV table called "CSVContent" with the following content:

Name;Category;Price
Apple;Fruit;0.99
Spaghetti;Pasta;5.49
Chicken Breast;Meat;8.99
Broccoli;Vegetable;2.49

To select every value in the "Name" column and account for the custom delimiter (;):

SELECT A.ID, X.Name FROM [TableWithCSVField] A CROSS APPLY CSVTABLE(A.CSVContent,';') WITH (Name VARCHAR(255)) AS X

-- Results:
|ID|Name|
-----------
|1|Apple|
|1|Spaghetti|
|1|Chicken Breast|
|1|Broccoli|

To create new records, use INSERT statements.

INSERT Syntax

The INSERT statement specifies the columns to be inserted and the new column values. You can specify the column values in a comma-separated list in the VALUES clause, as shown in the following example:

INSERT INTO <table_name> 
( <column_reference> [ , ... ] )
VALUES 
( { <expression> | NULL } [ , ... ] ) 
  

<expression> ::=
  | @ <parameter> 
  | ?
  | <literal>

INSERT INTO [CData].[Entities].Customers (CompanyName) VALUES ('Caterpillar')

To modify existing records, use UPDATE statements.

Update Syntax

The UPDATE statement takes as input a comma-separated list of columns and new column values as name-value pairs in the SET clause, as shown in the following example:

UPDATE <table_name> SET <select_statement> | {<column_reference> = <expression> [ , ... ]} WHERE { _id = <expression>  } [ { AND | OR } ... ] 

<expression> ::=
  | @ <parameter> 
  | ?
  | <literal>

The following is an example query:

UPDATE [CData].[Entities].Customers SET CompanyName='Caterpillar' WHERE _id = @my_id

To delete information from a table, use DELETE statements.

DELETE Syntax

The DELETE statement requires the table name in the FROM clause and the row's primary key in the WHERE clause, as shown in the following example:

<delete_statement> ::= DELETE FROM <table_name> WHERE { _id = <expression> } [ { AND | OR } ... ]

<expression> ::=
  | @ <parameter> 
  | ?
  | <literal>

The following is an example query:

DELETE FROM [CData].[Entities].Customers WHERE _id = @my_id

When caching is enabled, CACHE statements provide complete control over the data that is cached and the table to which it is cached. The CACHE statement executes the SELECT statement specified and caches its results to a table with the same name in the cache database or to table specified in <cached_table_name>. The connector updates or inserts rows to the cache depending on whether or not they already exist in the cache, so the primary key, which is used to identify existing rows, must be included in the selected columns.

See Caching Data for more information on different caching strategies.

CACHE Statement Syntax

The cache statement may include the following options that alter its behavior:

CACHE [ <cached_table_name> ] [ WITH TRUNCATE | AUTOCOMMIT | SCHEMA ONLY | DROP EXISTING | ALTER SCHEMA ] <select_statement> 

WITH TRUNCATE

If this option is set, the connector removes existing rows in the cache table before adding the selected rows. Use this option if you want to refresh the entire cache table but keep its existing schema.

AUTOCOMMIT

If this option is set, the connector commits each row individually. Use this option if you want to ignore the rows that could not be cached due to some reason. By default, the entire result set is cached as a single transaction.

DROP EXISTING

If this option is set, the connector drops the existing cache table before caching the new results. Use this option if you want to refresh the entire cache table, including its schema.

SCHEMA ONLY

If this option is set, the connector creates the cache table based on the SELECT statement without executing the query.

ALTER SCHEMA

If this option is set, the connector alters the schema of the existing table in the cache if it does not match the schema of the SELECT statement. This option results in new columns or dropped columns, if the schema of the SELECT statement does not match the cached table.

Common Queries

Use the following cache statement to cache all rows of a table:

CACHE SELECT * FROM [CData].[Entities].Customers

Use the following cache statement to cache all rows of a table into the cache table Cached[CData].[Entities].Customers:

CACHE CachedCustomers SELECT * FROM [CData].[Entities].Customers

Use the following cache statement for incremental caching. The DateModified column may not exist in all tables. The cache statement shows how incremental caching would work if there were such a column. Also, notice that, in this case, the WITH TRUNCATE and DROP EXISTING options are specifically omitted, which would have deleted all existing rows.

CACHE CachedCustomers SELECT * FROM [CData].[Entities].Customers WHERE DateModified > '2013-04-04'

Use the following cache statements to create a table with all available columns that will then cache only a few of them. The sequence of statements cache only City and CompanyName even though the cache table Cached[CData].[Entities].Customers has all the columns in [CData].[Entities].Customers.

CACHE CachedCustomers SCHEMA ONLY SELECT * FROM [CData].[Entities].Customers
CACHE CachedCustomers SELECT City, CompanyName FROM [CData].[Entities].Customers

To execute stored procedures, you can use EXECUTE or EXEC statements.

EXEC and EXECUTE assign stored procedure inputs, referenced by name, to values or parameter names.

Stored Procedure Syntax

To execute a stored procedure as an SQL statement, use the following syntax:

 
{ EXECUTE | EXEC } <stored_proc_name> 
{
  [ @ ] <input_name> = <expression>
} [ , ... ]

<expression> ::=
  | @ <parameter> 
  | ?
  | <literal>

Example Statements

Reference stored procedure inputs by name:

EXECUTE my_proc @second = 2, @first = 1, @third = 3;

Execute a parameterized stored procedure statement:

EXECUTE my_proc second = @p1, first = @p2, third = @p3; 

PIVOT and UNPIVOT can be used to change a table-valued expression into another table.

PIVOT

PIVOT Synax

 
"SELECT 'AverageCost' AS Cost_Sorted_By_Production_Days, [0], [1], [2], [3], [4]
FROM
(
SELECT DaysToManufacture, StandardCost
FROM Production.Product
) AS SourceTable
PIVOT
(
AVG(StandardCost)
FOR DaysToManufacture IN ([0], [1], [2], [3], [4])
) AS PivotTable;"

UNPIVOT

UNPIVOT Sytax

 
"SELECT VendorID, Employee, Orders
FROM
(SELECT VendorID, Emp1, Emp2, Emp3, Emp4, Emp5
FROM pvt) p
UNPIVOT
(Orders FOR Employee IN
(Emp1, Emp2, Emp3, Emp4, Emp5)
)AS unpvt;"

For further information on PIVOT and UNPIVOT, see FROM clause plus JOIN, APPLY, PIVOT (Transact-SQL)

Use INSERT INTO SELECT queries to select a list of records from one table and insert those same records into another table as a group. Inserting batches of records in this way may result in improved query performance compared to using many individual INSERT INTO queries.

The table whose records are selected for insertion into another table can be either a real table or a user-defined temporary table.

Inserting Records from Real Tables

To insert a group of records from one real, non-temporary, source table into another destination table, you can use an INSERT INTO SELECT query. This type of query is formatted similarly to a standard INSERT INTO query, except the VALUES clause is substituted with a SELECT query targeting the source table. All records matched by the embedded SELECT query are inserted into the destination table.

If the source table and destination tables have different column names, you must map columns from the source table to the corresponding columns in the destination table you want to insert them into. Perform this mapping by specifying the destination table columns in the same order as the source table columns you want to match them with. For example:

INSERT INTO DestinationTable (A,B,C,D) SELECT Q,R,S,T FROM SourceTable

In this example, the first source column (Q) is inserted into the first destination column (A), the second source column (R) is inserted into the second destination column (B), and so on.

If the source table and destination table both have the same column list with the same names, you can use a streamlined query.

INSERT INTO DestinationTableWithSameColumns SELECT * FROM SourceTable

In this example, there is no need to specify a list of columns for either the source or destination table, because their metadata already matches.

Inserting Records from Temporary Tables

You can manually define and populate temporary tables to hold a list of records for later bulk insertion.

Populate the Temporary Table

To create a temporary table, you must give it a name ending in "#TEMP" and execute an INSERT INTO query using that name, as if that table already existed in the database. After executing the first INSERT INTO, the temporary table exists and can receive subsequent INSERTs. For example:

INSERT INTO [CData].[Entities].Customers#TEMP (CompanyName, MyCustomField__c) VALUES ('New Customers', '9000');
INSERT INTO [CData].[Entities].Customers#TEMP (CompanyName, MyCustomField__c) VALUES ('New Customers 2', '9001');
INSERT INTO [CData].[Entities].Customers#TEMP (CompanyName, MyCustomField__c) VALUES ('New Customers 3', '9002');

This creates a temporary table called [CData].[Entities].Customers#TEMP with two columns and three rows of data. Since type cannot be determined on the temporary table itself, all values are considered strings and later converted to the proper type when they are inserted together into the real (non-temporary) table of interest.

Insert Temporary Table Contents into Real Tables

Once your temporary table is populated, execute an INSERT INTO SELECT query targeting the real (non-temporary) table you want to insert the temporary table's records into. This is formatted similarly to a standard INSERT INTO query, except the VALUES clause is substituted with a SELECT query targeting the matching columns in the temporary table. For example:

INSERT INTO [CData].[Entities].Customers (CompanyName, MyCustomField__c) SELECT CompanyName, MyCustomField__c FROM [CData].[Entities].Customers#TEMP

Results