Cross-Cutting Features¶
Features that span multiple additory functions.
Pipe Compatibility¶
Some functions work with df.pipe() for fluent method chaining. add.to(), add.synthetic(), and add.<dynamic>() are pipe-friendly. add.transform() and add.scan() are not (their first argument is a mode string, not a DataFrame).
Lineage Tracking¶
Track data provenance across operations with lineage=True. View the full operation history, column sources, and row mappings with add.scan('@lineage', df). Session-only by design — lineage lives in memory and is never written to disk.
Logging & Timing¶
Enable operation logging with logging=True or get detailed timing breakdowns with the ADDITORY_TIMING environment variable. Useful for debugging and performance profiling.
Type Handling¶
Additory works with both pandas and polars DataFrames. Output type matches input type by default. Use as_type to force a specific output type. Polars is the native internal format; pandas inputs are converted via Arrow IPC.