Release v0.5.1

Fix bug to properly resolve initialization script paths starting with ~ (i.e. user's home dir) supplied to --init-path.

Release v0.5.0

Add support for initialization scripts. By default harlequin-databricks will attempt to run an initialization script of SQL commands against Databricks from ~/.databricksrc or from the file specified via the --init-path CLI option. This means you can e.g. set a default catalog (USE CATALOG ...), timezone, or set of user-defined variables for the session. It is possible to disable initialization via the --no-init CLI flag, having no ~/.databricksrc file, or feeding a non-existent file path to --init-path. (#14)

Release v0.4.0

Features

• Add support for cancelling queries mid-flight. Requires Harlequin >=1.24.0 which introduced the "Cancel Query" button.
• Add support for Azure in OAuth user-to-machine authentication for Databricks running on Azure.
• Add support for OAuth machine-to-machine (M2M) authentication to Databricks. This allows you to use service principle credentials to connect to Databricks via Harlequin, useful for testing. To use OAuth M2M, supply a --client-id and a --client-secret (i.e. a service principle OAuth token) via CLI arguments.
• Better error handling and debug messages when database query errors occur.

Release v0.3.1

Bug Fixes

• Fix UnicodeDecodeError on Windows due to incorrectly attempting to read functions.csv using
  CP-1252. Now UTF-8 is enforced on all file writes and reads (#7).

Features

• Update list of Databricks SQL functions for completions (valid as of August 2, 2024).
• Add harlequin-databricks screenshot to README.

Release v0.3.0

Features

• Implement the close() method of HarlequinConnection so that the Databricks connection is
  closed when Harlequin exits.
• Update list of Databricks SQL functions for completions (valid as of April 24, 2024).

Release v0.2.1

Bug Fixes

• Fix SQL created in the Query Editor by double clicking on catalog, schema, table or column
  names in the Data Catalog pane. Previously items were wrapped in double quotes, which is invalid
  Databricks SparkSQL.
• Fix outstanding ruff and mypy errors.

Release v0.2.0

Features

• Faster indexing for Unity Catalog assets. Makes loading & refreshing Harlequin's Data Catalog pane much quicker for Databricks instances running Unity Catalog. The new method fetches metadata for Unity Catalog assets from system.information_schema, requiring only two SQL queries to do so. This is much faster than the old method, which required separate SQL calls for each table. The old method is still in use for indexing legacy metastores (e.g. hive_metastore) however, as that metadata is not contained in Information schema: https://docs.databricks.com/en/sql/language-manual/sql-ref-information-schema.html
• Add new command line flag --skip-legacy-indexing to skip the indexing of legacy metastores. Setting this flag is recommended if your Databricks instance runs Unity Catalog and you do not want the overhead of slow indexing of legacy metastores. When this flag is set, only Unity Catalog assets will show up in the Data Catalog pane.

Bug Fixes

• Fix minor formatting (black), types (mypy) & linting (ruff) issues, and a test failure.
• Don't use underscores in CLI option names, i.e. make it so the only acceptable option version is the one written with hyphens not underscores.

Release v0.1.1

• Ensure catalogs, schemas & tables are sorted alphabetically in Data Catalog pane.
• Ensure order of columns shown in Data Catalog pane is correct for each table.

Release v0.1.0

Adds a Harlequin adapter for Databricks SQL warehouses and DBR interactive clusters.

Available on PyPI and installable via:

pip install harlequin-databricks