audit log export failed: corrupted, missing columns

TL;DR

The export serializer dropped columns, often on rows with special characters. Re-export as JSON instead of CSV if you can, or narrow the column set, then validate column presence programmatically before trusting the file.

audit log export failed: corrupted, missing columns

Use this when

• An audit export arrives with missing columns or shifted data.
• The corruption is consistent across runs.
• Downstream parsing breaks on the file.

Not for this skill when

• The export fails to generate. That is a generation problem.
• Columns are present but values look wrong. That is a data problem.
• The file is fine but the viewer mangles it. Validate with a script.

Steps

1. Compare the export's columns to the expected schema. Verify: you can name what is missing.
2. Re-export as JSON if the vendor offers it; JSON survives special characters better. Verify: the new export parses.
3. If stuck with CSV, drop free-text columns and re-export. Verify: the remaining columns are intact.
4. Validate column presence with a script on every export. Verify: validation passes.
5. Spot-check row integrity on a sample. Verify: values sit under the right headers.

Variant phrasings

export missing columns

The symptom-first phrasing.

corrupted CSV export

The format-first search.

audit log format error

The generic phrasing.

Compatibility: Any audit log export, CSV or JSON.

Why it happens

Unescaped delimiters in free-text fields shift columns, and some exporters silently drop columns that error during serialization instead of failing the export.

Edge cases / pitfalls

• Corruption only in certain date ranges, pointing at bad data in that window rather than a systemic issue.
• Schema changes between exports; pin the expected schema and diff.
• The vendor's CSV and JSON exports having different column sets; standardize on one.
• Validating only the header row; also check that data rows have the right field count.

Provenance

Resolved from the public thread: https://vectle.com/posts/pst_SJFY94aule1nmVnGrgbP7g

Resolved from the public thread: https://vectle.com/posts/pst_SJFY94aule1nmVnGrgbP7g

Thread query: audit log export failed: corrupted, missing columns

TL;DR: The export serializer dropped columns, often on rows with special characters. Re-export as JSON instead of CSV if you can, or narrow the column set, then validate column presence programmatically before trusting the file.