v0.2.1

• Fixed row converter to properly handle nullable values (#53). In this release, the row converter in the databricks.labs.lsql.core module has been updated to handle nullable values correctly. A new method StatementExecutionExt has been added, which manages the handling of nullable values during SQL statement execution. The Row class has also been modified to include nullable values, improving the robustness and flexibility of SQL execution in dealing with various data types, including null values. These enhancements increase the overall reliability of the system, making it more production-ready.
• Improved integration test coverage (#52). In this release, the project's integration test coverage has been significantly improved through several changes. A new function, make_random(), has been added to the conftest.py file to generate a random string of fixed length, aiding in the creation of more meaningful and readable random strings for integration tests. A new file, test_deployment.py, has been introduced, containing a test function for deploying a database schema and verifying successful data retrieval via a view. The test_integration.py file has been renamed to test_core.py, with updates to the test_fetch_one function to test the fetch_one method using a SQL query with an aliased value. Additionally, a new Foo dataclass has been added to the tests/integration/views/__init__.py file, supporting integration test coverage. Lastly, a new SQL query has been added to the integration test suite, located in the some.sql file, which retrieves data from a table named foo in the inventory schema. These changes aim to enhance the overall stability, reliability, and coverage of the project's integration tests. Note: The changes to the .gitignore file and the improvements to the StatementExecutionBackend class in the backends.py file are not included in this summary, as they were described in the opening statement.
• Rely on hatch being present on the build machine (#54). In this release, we have made significant changes to how we manage our build process and toolchain configuration. We have removed the need to manually install hatch version 1.7.0 in the build machine, and instead, rely on its presence, adding it to the list of required tools in the toolchain configuration. The command to create a virtual environment using hatch has also been added, and the pre_setup section no longer includes installing hatch, assuming its availability. We have also updated the hatch package version from 1.7.0 to 1.9.4, which may include bug fixes, performance improvements, or new features. This change may impact the behavior of any existing functionality that relies on hatch. The pyproject.toml file has been modified to update the fmt and verify sections, with ruff check . --fix replacing ruff . --fix and the removal of black --check . and isort . --check-only. A new configuration for isort has also been added to specify the databricks.labs.blueprint package as a known first-party package, enabling more precise management of imports related to that package. These changes simplify the build process and ensure that the project is using a more recent version of the hatch package for packaging and distributing Python projects.
• Updated sqlglot requirement from ~=22.3.1 to >=22.3.1,<22.5.0 (#51). In this release, we have updated the version constraint for the sqlglot package in our project's pyproject.toml file. Previously, we had set the constraint to ~=22.3.1, allowing for any version with the same major and minor numbers but different patch numbers. With this update, we have changed the constraint to >=22.3.1,<22.5.0. This change enables our project to utilize bug fixes and improvements made in the latest patch versions of sqlglot, while still preventing it from inadvertently using any breaking changes introduced in version 22.5.0 or later versions. This modification allows us to take advantage of the latest features and improvements in sqlglot while maintaining compatibility and stability in our project.

Dependency updates:

• Updated sqlglot requirement from ~=22.3.1 to >=22.3.1,<22.5.0 (#51).

Contributors: @nfx, @dependabot[bot]

v0.2.0

• Added MockBackend.rows("col1", "col2")[(...), (...)] helper (#49). In this release, we have added a new helper method MockBackend.rows("col1", "col2")[(...), (...)] to simplify testing with MockBackend. This method allows for the creation of rows using a more concise syntax, taking in the column names and a list of values to be used for each column, and returning a list of Row objects with the specified columns and values. Additionally, a __eq__ method has been introduced to check if two rows are equal by converting the rows to dictionaries using the existing as_dict method and comparing them. The __contains__ method has also been modified to improve the behavior of the in keyword when used with rows, ensuring columns can be checked for membership in the row in a more intuitive and predictable manner. These changes make it easier to test and work with MockBackend, improving overall quality and maintainability of the project.

Contributors: @nfx

v0.1.1

• Updated project metadata (#46). In this release, the project metadata has been updated to reflect changes in the library's capabilities and dependencies. The project now supports lightweight SQL statement execution using the Databricks SDK for Python, setting it apart from other solutions. The library size comparison in the documentation has been updated, reflecting an increase in the compressed and uncompressed size of Databricks Labs LightSQL, as well as the addition of a new direct dependency, SQLglot. The project's dependencies and URLs in the pyproject.toml file have also been updated, including a version update for databricks-labs-blueprint and the removal of a specific range for PyYAML.

Dependency updates:

• Updated sqlglot requirement from ~=22.2.1 to ~=22.3.1 (#43).

Contributors: @dependabot[bot], @nfx

v0.1.0

• Ported StatementExecutionExt from UCX (#31).

Contributors: @dependabot[bot], @nfx