CSV Repair Kit

Runs in your browser

CSV Column Mismatch Finder

Find rows with the wrong number of fields and pad or trim them.

Your files never leave your device.

Drop a CSV with jagged rows

CSV, TSV, or text. Processed locally — the file is never uploaded.

Try a sample:

Why rows change width

Importers that \u201csplit on comma\u201d without a real CSV parser will also create jagged files when they write back out. So will truncated FTP drops and unquoted free-text comments. Databases then reject the load.

What this finder reports

Each row whose field count differs from the mode is listed with the observed width. Repair can pad missing fields with empties or drop extras on long rows. Preview highlights those rows so you can confirm you are not deleting a real column.

Example

id,name,qty
1,Nails,10
2,Hammer,1,warehouse
3,Tape

Padded/trimmed to three columns:

id,name,qty
1,Nails,10
2,Hammer,1
3,Tape,

Limitations

Extra fields are not intelligently merged. If warehouse belongs in a fourth header, add that column in a spreadsheet first. Quote errors that swallowed later rows should be fixed with the quote tool before aligning columns.

FAQ

How is the expected width chosen?

The most common column count across rows (the statistical mode). The header is not assumed to be truth if most data rows disagree.

Should I pad or trim?

Pad when a field was omitted. Trim when an unquoted comma created extra fields you cannot reconstruct automatically.

What about multiline quoted fields?

They count as one field. A mismatch is a different number of fields, not extra physical lines.

Related tools