Databricks SQL Driver for Node.js

The Databricks SQL Driver for Node.js is a Node.js library that allows you to use JavaScript code to run SQL commands on Databricks compute resources.

Requirements

• A development machine running Node.js, version 14 or higher. To print the installed version of Node.js, run the command node -v. To install and use different versions of Node.js, you can use tools such as Node Version Manager (nvm).

• Node Package Manager (npm). Later versions of Node.js already include npm. To check whether npm is installed, run the command npm -v. To install npm if needed, you can follow instructions such as the ones at Download and install npm.

• The @databricks/sql package from npm. To install the @databricks/sql package in your Node.js project as a dependency, use npm to run the following command from within the same directory as your project:

  npm i @databricks/sql
  

• If you want to install and use TypeScript in your Node.js project as devDependencies, use npm to run the following commands from within the same directory as your project:

  npm i -D typescript
  npm i -D @types/node
  

• An existing cluster or SQL warehouse.

• The Server Hostname and HTTP Path value for the existing cluster or SQL warehouse.

  • Get these values for a cluster.

  • Get these values for a SQL warehouse.

• A Databricks personal access token.

  • Manage personal access tokens.

Specify the connection variables

To access your cluster or SQL warehouse, the Databricks SQL Driver for Node.js uses connection variables named token, server_hostname and http_path, representing your Databricks personal access token and your cluster’s or SQL warehouse’s Server Hostname and HTTP Path values, respectively.

The Databricks personal access token value for token is similar to the following: dapi12345678901234567890123456789012.

The Server Hostname value for server_hostname is similar to the following: dbc-a1b2345c-d6e7.cloud.databricks.com.

The HTTP Path value for http_path is similar to the following: for a cluster, sql/protocolv1/o/1234567890123456/1234-567890-abcdefgh; and for a SQL warehouse, /sql/1.0/endpoints/a1b234c5678901d2.

Note

As a security best practice, you should not hard code these connection variable values into your code. Instead, you should retrieve these connection variable values from a secure location. For example, the code example later in this article uses environment variables.

Query data

The following code example demonstrates how to call the Databricks SQL Driver for Node.js to run a basic SQL query on a Databricks compute resource. This command returns the first two rows from the diamonds table.

The diamonds table is included in Sample datasets. This table is also featured in Tutorial: Query data with notebooks.

This code example retrieves the token, server_hostname and http_path connection variable values from a set of Databricks environment variables. These environment variables have the following environment variable names:

• DATABRICKS_TOKEN, which represents your Databricks personal access token from the requirements.

• DATABRICKS_SERVER_HOSTNAME, which represents the Server Hostname value from the requirements.

• DATABRICKS_HTTP_PATH, which represents the HTTP Path value from the requirements.

You can use other approaches to retrieving these connection variable values. Using environment variables is just one approach among many.

The following code example demonstrates how to call the Databricks SQL Connector for Node.js to run a basic SQL command on a cluster or SQL warehouse. This command returns the first two rows from the diamonds table.

const { DBSQLClient } = require('@databricks/sql');

const token          = process.env.DATABRICKS_TOKEN;
const serverHostname = process.env.DATABRICKS_SERVER_HOSTNAME;
const httpPath       = process.env.DATABRICKS_HTTP_PATH;

if (!token || !serverHostname || !httpPath) {
  throw new Error("Cannot find Server Hostname, HTTP Path, or personal access token. " +
                  "Check the environment variables DATABRICKS_TOKEN, " +
                  "DATABRICKS_SERVER_HOSTNAME, and DATABRICKS_HTTP_PATH.");
}

const client = new DBSQLClient();
const connectOptions = {
  token: token,
  host: serverHostname,
  path: httpPath
};

client.connect(connectOptions)
  .then(async client => {
    const session = await client.openSession();
    const queryOperation = await session.executeStatement(
      'SELECT * FROM default.diamonds LIMIT 2',
      {
        runAsync: true,
        maxRows: 10000 // This option enables the direct results feature.
      }
    );

    const result = await queryOperation.fetchAll();

    await queryOperation.close();

    console.table(result);

    await session.close();
    await client.close();
})
.catch((error) => {
  console.log(error);
});

import { DBSQLClient } from '@databricks/sql';
import IDBSQLSession from '@databricks/sql/dist/contracts/IDBSQLSession';
import IOperation from '@databricks/sql/dist/contracts/IOperation';

const serverHostname: string = process.env.DATABRICKS_SERVER_HOSTNAME || '';
const httpPath: string       = process.env.DATABRICKS_HTTP_PATH || '';
const token: string          = process.env.DATABRICKS_TOKEN || '';

if (serverHostname == '' || httpPath == '' || token == '') {
  throw new Error("Cannot find Server Hostname, HTTP Path, or personal access token. " +
                  "Check the environment variables DATABRICKS_SERVER_HOSTNAME, " +
                  "DATABRICKS_HTTP_PATH, and DATABRICKS_TOKEN.");
}

const client: DBSQLClient = new DBSQLClient();
const connectOptions = {
  host: serverHostname,
  path: httpPath,
  token: token
};

client.connect(connectOptions)
  .then(async client => {
    const session: IDBSQLSession = await client.openSession();

    const queryOperation: IOperation = await session.executeStatement(
      'SELECT * FROM default.diamonds LIMIT 2',
      {
        runAsync: true,
        maxRows: 10000 // This option enables the direct results feature.
      }
    );

    const result = await queryOperation.fetchAll();

    await queryOperation.close();

    console.table(result);

    await session.close();
    client.close();
  })
  .catch((error) => {
    console.log(error);
});

Output:

┌─────────┬─────┬────────┬───────────┬───────┬─────────┬────────┬───────┬───────┬────────┬────────┬────────┐
│ (index) │ _c0 │ carat  │    cut    │ color │ clarity │ depth  │ table │ price │   x    │   y    │   z    │
├─────────┼─────┼────────┼───────────┼───────┼─────────┼────────┼───────┼───────┼────────┼────────┼────────┤
│    0    │ '1' │ '0.23' │  'Ideal'  │  'E'  │  'SI2'  │ '61.5' │ '55'  │ '326' │ '3.95' │ '3.98' │ '2.43' │
│    1    │ '2' │ '0.21' │ 'Premium' │  'E'  │  'SI1'  │ '59.8' │ '61'  │ '326' │ '3.89' │ '3.84' │ '2.31' │
└─────────┴─────┴────────┴───────────┴───────┴─────────┴────────┴───────┴───────┴────────┴────────┴────────┘

Sessions

All IDBSQLSession methods that return IOperation objects in the API Reference have the following common parameters that affect their behavior:

• Setting runAsync to true starts asynchronous mode. IDBSQLSession methods put operations into the queue and return as quickly as possible. The current state of the returned IOperation object might vary, and the client is responsible for checking its status before using the returned IOperation. See Operations. Setting runAsync to false means that IDBSQLSession methods wait for operations to complete. Databricks recommends always setting runAsync to true.

• Setting maxRows to a non-null value enables direct results. With direct results, the server tries to wait for operations to complete and then fetches a portion of the data. Depending on how much work the server was able to complete within the defined time, IOperation objects return in some intermediate state instead of in some pending state. Very often all the metadata and query results are returned within a single request to the server. The server uses maxRows to determine how many records it can return immediately. However, the actual chunk may be of a different size; see IDBSQLSession.fetchChunk. Direct results are enabled by default. Databricks recommends against disabling direct results.

Operations

As described in Sessions, IOperation objects that are returned by IDBSQLSession session methods in the API Reference are not fully populated. The related server operation might still be in progress, such as waiting for the Databricks SQL warehouse to start, running the query, or fetching the data. The IOperation class hides these details from users. For example, methods such fetchAll, fetchChunk, and getSchema wait internally for operations to complete and then return results. You can use the IOperation.finished() method to explicitly wait for operations to complete. These methods take a callback that is periodically called while waiting for operations to complete. Setting the progress option to true attempts to request extra progress data from the server and pass it to that callback.

The close and cancel methods can be called at any time. When called, they immediately invalidate the IOperation object; all pending calls such as fetchAll, fetchChunk, and getSchema are immediately canceled and an error is returned. In some cases, the server operation might have already completed and the cancel method affects only the client.

The fetchAll method calls fetchChunk internally and collects all of the data into an array. While this is convenient, it might cause out of memory errors when used on large datasets. fetchAll options are typically passed to fetchChunk.

Fetch chunks of data

Fetching data chunks uses the following code pattern:

do {
  const chunk = await operation.fetchChunk();
  // Process the data chunk.
} while (await operation.hasMoreRows());

The fetchChunk method in the API Reference processes data in small portions to reduce memory consumption. fetchChunk first waits for operations to complete if they have not already completed, then calls a callback during the wait cycle, and then fetches the next data chunk.

You can use the maxRows option to specify the desired chunk size. However, the returned chunk might have a different size, smaller or even sometimes larger. fetchChunk does not attempt to prefetch data internally, in order to slice it into the requested portions. It sends the maxRows option to then server and returns whatever the server returns. Do not confuse this maxRows option with the one in IDBSQLSession. maxRows passed to fetchChunk defines the size of each chunk and does not do anything else.

Configure logging

The logger provides information for debugging problems with the connector. All DBSQLClient objects are instantiated with a logger that prints to the console, but by passing in a custom logger, you can send this information to a file. The following example shows how to configure a logger and change its level.

const { DBSQLLogger, LogLevel } = require('@databricks/sql');
const logger = new DBSQLLogger({
  filepath: 'log.txt',
  level: LogLevel.info
});

// Set logger to different level.
logger.setLevel(LogLevel.debug);

import { DBSQLLogger, LogLevel } from '@databricks/sql';
const logger = new DBSQLLogger({
  filepath: 'log.txt',
  level: LogLevel.info,
});

// Set logger to different level.
logger.setLevel(LogLevel.debug);

For additional examples, see the examples folder in the databricks/databricks-sql-nodejs repository on GitHub.

Additional resources

• The Databricks SQL Driver for Node.js repository on GitHub

• Getting started with the Databricks SQL Driver for Node.js

• Troubleshooting the Databricks SQL Driver for Node.js

API reference

Classes

DBSQLClient class

Main entry point for interacting with a database.

Methods

connect method

Opens a connection to the database.

Parameters
options Type: ConnectionOptions The set of options used to connect to the database. The host, path, and token fields must be populated. Example: const client: DBSQLClient = new DBSQLClient(); client.connect( { host: serverHostname, path: httpPath, token: token } )

Returns: Promise<IDBSQLClient>

openSession method

Opens session between DBSQLClient and database.

Parameters
request Type: OpenSessionRequest A set of optional parameters for specifying initial schema and initial catalog Example: const session = await client.openSession( {initialCatalog: 'catalog'} );

Returns: Promise<IDBSQLSession>

getClient method

Returns internal thrift TCLIService.Client object. Must be called after DBSQLClient has connected.

No parameters

Returns TCLIService.Client

close method

Closes the connection to the database and releases all associated resources on the server. Any additional calls to this client will throw error.

No parameters.

No return value.

DBSQLSession class

DBSQLSessions are primarily used for the execution of statements against the databbase as well as various metadata fetching operations.

Methods

executeStatement method

Executes a statement with the options provided.

Parameters
statement Type: str The statement to be executed.
options Type: ExecuteStatementOptions A set of optional parameters for determining query timeout, max rows for direct results, and whether to run the query asynchronously. By default maxRows is set to 10000. If maxRows is set to null, the operation will run with the direct results feature off. Example: const session = await client.openSession( {initialCatalog: 'catalog'} ); queryOperation = await session.executeStatement( 'SELECT "Hello, World!"', { runAsync: true } );

Returns: Promise<IOperation>

close method

Closes the session. Must be done after using session.

No parameters.

No return value.

getId method

Returns the GUID of the session.

No parameters.

Returns: str

getTypeInfo method

Returns information about supported data types.

Parameters
request Type: TypeInfoRequest Request parameters.

Returns: Promise<IOperation>

getCatalogs method

Gets list of catalogs.

Parameters
request Type: CatalogsRequest Request parameters.

Returns: Promise<IOperation>

getSchemas method

Gets list of schemas.

Parameters
request Type: SchemasRequest Request parameters. Fields catalogName and schemaName can be used for filtering purposes.

Returns: Promise<IOperation>

getTables method

Gets list of tables.

Parameters
request Type: TablesRequest Request parameters. Fields catalogName and schemaName and tableName can be used for filtering.

Returns: Promise<IOperation>

getFunctions method

Gets list of tables.

Parameters
request Type: FunctionsRequest Request parameters. Field functionName is required.

Returns: Promise<IOperation>

getPrimaryKeys method

Gets list of primary keys.

Parameters
request Type: PrimaryKeysRequest Request parameters. Fields schemaName and tableName are required.

Returns: Promise<IOperation>

getCrossReference method

Gets information about foreign keys between two tables.

Parameters
request Type: CrossReferenceRequest Request parameters. Schema, Parent, and Catalog name must be specified for both tables.

Returns: Promise<IOperation>

DBSQLOperation class

DBSQLOperations are created by DBSQLSessions and can be used to fetch the results of statements and check up on their execution. Data is fetched through functions fetchChunk and fetchAll.

Methods

getId method

Returns the GUID of the operation.

No parameters.

Returns: str

fetchAll method

Waits for operation completion, then fetches all rows from operation.

Parameters: None

Returns: Promise<Array<object>>

fetchChunk method

Waits for operation completion, then fetches up to a specified number of rows from an operation.

Parameters
options Type: FetchOptions Options used to fetch. Currently, the only option is maxRows, which corresponds to the max number of data objects to be returned in any given array.

Returns: Promise<Array<object>>

close method

Closes the operation and releases all associated resources. Must be done after no longer using operation.

No parameters.

No return value.