Export and Transformation Log

Unit ID: M09-U09 Estimated active time: 25-35 minutes

Write clean output separately from raw input and make the schema predictable.

clean.to_csv(output_path, index=False, encoding="utf-8", lineterminator="\n")

The transformation log should name source files and checksums, row counts, rejected reasons, cleaning rules, join results, output columns, and output file names. Deterministic learning outputs should not include a changing timestamp unless time is part of the lesson.

Read the exported file back and check its columns, row count, and dtypes. Export success alone does not prove the result is correct.

Practice: write and reload the clean table, then compare its required columns and learner-module key count.