Release v0.6.2

Bug Fixes

• Ensure LICENSE file gets packaged in distribution - this is needed for the conda-forge feedstock (cf. conda-forge/harlequin-databricks-feedstock#13) (#22).

Release v0.6.1

1. Features

• Relax harlequin version to >=2.0.4 to allow conda-forge builds again (#21).
  • harlequin's conda-forge feedstocks for >=2.0.5 are currently blocked by tree-sitter-sql not yet being on conda-forge. This is then blocked harlequin-databricks feedstock for 0.6.0
    conda-forge/harlequin-feedstock#56
    conda-forge/harlequin-databricks-feedstock#12

2. Code Quality

• Add classifiers and keywords to pyproject.toml (#20).
• Update URLs for Databricks SQL keywords (#20).
• Reorder deps in pyproject.toml (#21).

Release v0.6.0

1. Breaking Changes

• Require databricks-sql-connector>=4.0.4 to fix unable to index data catalog bug (#18).
• Require harlequin>=2.1.2.

2. Features

• Recommend uv as installation method.
• Update functions syntax highlighting list.

3. Code Quality

• Use uv for package management, build backend and CI.
• Use ruff for linting and code formatting.
• Drop functional tests in CI against Databricks as test Databricks instance no longer works.

Release v0.5.2

Breaking Changes

• Drops support for Python 3.8 (#17).
• Require latest versions of Harlequin >=1.25.2, databricks-sql-connector >=3.7.0 and
  databricks-sdk >=0.40.0 (optional extra), and avoid upper version bounds on requirements to avoid
  dependency hell (more here: databricks/databricks-sql-python#452 &
  tconbeer/harlequin@a50dfac#diff-50c86b7ed8ac2cf95bd48334961bf0530cdc77b5a56f852c5c61b89d735fd711R64)
  (#17).

Code Quality

• Reduce unused mypy ignores (#17).

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.