v0.9.4 (2026-03-31)

What’s new

Added

• reader_kwargs attribute to PipelineNodeDataSource

Fixed

• TableDataSink.exists() method.
• Removed logs deduplication in Databricks notebooks
• Serialization roundtrip when schema_definition of BaseDataSink is used
• Roundtrip when creating a DataFrameSchema instance from a native dataframe schema.
• Provisioning laktory_context value when used inside a custom write used in a streaming query

Updated

• Clarified documentation by removing some base class attributes from child classes [#443]
• Clarified documentation of VariableType [#444]

Full Changelog