Databricks

Databricks on your own workspace: run SQL on a SQL warehouse, start and check jobs and runs, inspect and start or terminate clusters, browse Unity Catalog catalogs, schemas and tables, and list workspace folders. You give your workspace address and a personal access token.

15 actions API key 11 read2 write2 destructive Calls www.databricks.com

Connect

TypeScript

const conn = await setu.connections.initiate({ userId: "user_1", toolkit: "databricks" });
console.log(conn.connectUrl); // show this to your user
await conn.waitUntilActive();

Permissions requested

This toolkit uses a single credential and requests no scopes.

Actions

Execute SQL statementdestructivedatabricks.execute_sqlRun a SQL statement on a Databricks SQL warehouse and return up to the first rows inline. The statement can be anything the token's user may run, including INSERT, UPDATE, DELETE, DROP and TRUNCATE, so it is treated as destructive and needs confirmation: prefer plain SELECT statements and use row_limit. If it does not finish within wait_timeout it keeps running and you poll get_statement with the returned statement_id.

Input	Type		Description
`statement`	string	Required	The SQL to run (up to 16 MiB)
`warehouse_id`	string	Required	ID of the SQL warehouse that runs it, from list_warehouses
`catalog`	string	Optional	Default catalog for unqualified names (like USE CATALOG)
`schema`	string	Optional	Default schema for unqualified names (like USE SCHEMA)
`wait_timeout`	string	Optional	How long to wait for the result: 0s (return immediately) or 5s to 50s. Default 10s
`row_limit`	integer	Optional	Cap the number of rows returned (sets truncated when hit)
`parameters`	array of object	Optional	Named parameter markers (:name) used in the statement

TypeScript

await setu.execute({
  userId: "user_1",
  action: "databricks.execute_sql",
  input: {
    "statement": "<statement>",
    "warehouse_id": "<warehouse_id>"
  },
  confirm: true,
});

Get clusterreaddatabricks.get_clusterGet one cluster's state, Spark version, node type, size, auto-termination and creator.

Input	Type		Description
`cluster_id`	string	Required	Cluster ID, from list_clusters

TypeScript

await setu.execute({
  userId: "user_1",
  action: "databricks.get_cluster",
  input: {
    "cluster_id": "<cluster_id>"
  },
});

Get jobreaddatabricks.get_jobGet one job's definition: name, schedule, concurrency limit and its tasks.

Input	Type		Description
`job_id`	integer	Required	Job ID, from list_jobs

TypeScript

await setu.execute({
  userId: "user_1",
  action: "databricks.get_job",
  input: {
    "job_id": 1
  },
});

Get job runreaddatabricks.get_runGet the status of a job run: lifecycle state (PENDING, RUNNING, TERMINATED...), result state (SUCCESS, FAILED...), timings and a link to the run page.

Input	Type		Description
`run_id`	integer	Required	Run ID, from run_job

TypeScript

await setu.execute({
  userId: "user_1",
  action: "databricks.get_run",
  input: {
    "run_id": 1
  },
});

Get SQL statement resultreaddatabricks.get_statementCheck the status of a SQL statement started with execute_sql and, once SUCCEEDED, get the first chunk of its rows. Statements are removed about 12 hours after they finish.

Input	Type		Description
`statement_id`	string	Required	Statement ID returned by execute_sql

TypeScript

await setu.execute({
  userId: "user_1",
  action: "databricks.get_statement",
  input: {
    "statement_id": "<statement_id>"
  },
});

List Unity Catalog catalogsreaddatabricks.list_catalogsList the Unity Catalog catalogs the token's user can see, with owner, comment and type.

Input	Type		Description
`max_results`	integer	Optional	Maximum catalogs to return
`page_token`	string	Optional	Token from a previous call's next_page_token, for the next page

TypeScript

await setu.execute({
  userId: "user_1",
  action: "databricks.list_catalogs",
  input: {},
});

List clustersreaddatabricks.list_clustersList the all-purpose and job clusters in the workspace with ID, name, state (RUNNING, TERMINATED, PENDING...), Spark version, node type and worker count.

Input	Type		Description
`page_size`	integer	Optional	Clusters per page (max 100)
`page_token`	string	Optional	Token from a previous call's next_page_token, for the next page

TypeScript

await setu.execute({
  userId: "user_1",
  action: "databricks.list_clusters",
  input: {},
});

List jobsreaddatabricks.list_jobsList Databricks jobs (Lakeflow Jobs / workflows) in the workspace with their ID, name, creator and schedule. Optionally filter by exact name.

Input	Type		Description
`name`	string	Optional	Only jobs with exactly this name (case-insensitive)
`limit`	integer	Optional	Jobs per page (1-100, default 20)
`page_token`	string	Optional	Token from a previous call's next_page_token, for the next page

TypeScript

await setu.execute({
  userId: "user_1",
  action: "databricks.list_jobs",
  input: {},
});

List Unity Catalog schemasreaddatabricks.list_schemasList the schemas (databases) inside one Unity Catalog catalog.

Input	Type		Description
`catalog_name`	string	Required	Catalog name, from list_catalogs
`max_results`	integer	Optional	Maximum schemas to return
`page_token`	string	Optional	Token from a previous call's next_page_token, for the next page

TypeScript

await setu.execute({
  userId: "user_1",
  action: "databricks.list_schemas",
  input: {
    "catalog_name": "<catalog_name>"
  },
});

List Unity Catalog tablesreaddatabricks.list_tablesList the tables and views in one Unity Catalog schema. Set omit_columns to true for a short answer; leave it off to also get each table's column names and types.

Input	Type		Description
`catalog_name`	string	Required	Catalog name
`schema_name`	string	Required	Schema name, from list_schemas
`omit_columns`	boolean	Optional	Leave out the column lists (shorter output)
`max_results`	integer	Optional	Maximum tables to return
`page_token`	string	Optional	Token from a previous call's next_page_token, for the next page

TypeScript

await setu.execute({
  userId: "user_1",
  action: "databricks.list_tables",
  input: {
    "catalog_name": "<catalog_name>",
    "schema_name": "<schema_name>"
  },
});

List SQL warehousesreaddatabricks.list_warehousesList the SQL warehouses the token's user can use, with ID, name, size, state (STARTING, RUNNING, STOPPED...) and type. The warehouse ID is needed by execute_sql.

Input	Type		Description
`page_size`	integer	Optional	Maximum warehouses to return
`page_token`	string	Optional	Token from a previous call's next_page_token, for the next page

TypeScript

await setu.execute({
  userId: "user_1",
  action: "databricks.list_warehouses",
  input: {},
});

List workspace objectsreaddatabricks.list_workspace_objectsList the notebooks, folders, files and libraries directly inside a workspace folder path such as /Users/[email protected] or /Workspace/Shared.

Input	Type		Description
`path`	string	Required	Absolute workspace path of the folder, e.g. /Users/[email protected]

TypeScript

await setu.execute({
  userId: "user_1",
  action: "databricks.list_workspace_objects",
  input: {
    "path": "<path>"
  },
});

Run job nowwritedatabricks.run_jobStart a run of an existing job right now, optionally with job parameters. Starts compute and executes the job's tasks, which can change data. Use an idempotency_token to avoid starting the same run twice on retries. Poll get_run with the returned run_id.

Input	Type		Description
`job_id`	integer	Required	Job ID, from list_jobs
`job_parameters`	object	Optional	Job-level parameters as name to text value pairs
`idempotency_token`	string	Optional	Optional unique token (max 64 characters); a retry with the same token returns the same run

TypeScript

await setu.execute({
  userId: "user_1",
  action: "databricks.run_job",
  input: {
    "job_id": 1
  },
});

Start clusterwritedatabricks.start_clusterStart a terminated cluster. It begins using compute that is billed. The call returns at once; poll get_cluster until its state is RUNNING.

Input	Type		Description
`cluster_id`	string	Required	ID of the terminated cluster to start

TypeScript

await setu.execute({
  userId: "user_1",
  action: "databricks.start_cluster",
  input: {
    "cluster_id": "<cluster_id>"
  },
});

Terminate clusterdestructivedatabricks.terminate_clusterTerminate a cluster: it stops, running jobs and notebooks on it are interrupted, and its configuration is kept so it can be started again (this is not a permanent delete).

Input	Type		Description
`cluster_id`	string	Required	ID of the cluster to terminate

TypeScript

await setu.execute({
  userId: "user_1",
  action: "databricks.terminate_cluster",
  input: {
    "cluster_id": "<cluster_id>"
  },
  confirm: true,
});