TableMender

Your files stay in this browser

Advertising and analytics are currently off. We remember only a dismissed-notice preference for this tab. Cloudflare may use technical security cookies. Cookies & storage details · File privacy

Home / Tools
Make deliberate changes

CSV clean: duplicates, empty records & whitespace

Choose cleaning rules, review their impact, and export a new CSV. Keep the first duplicate key and preserve your source.

YOUR WORKBENCHLocal processing
  1. 01 Add data
  2. 02 Review
  3. 03 Export

Start with your source.

CSV/text only: 5 MiB, 20,000 records, 256 columns, 250,000 cells. Processing stays in your browser. Keep an original copy.

Auto detection needs confirmation. An uncertain separator blocks export; choose one explicitly.

Encoding applies to selected files. Pasted text is already decoded by your browser.

Choose cleaning rules

All rules start off. Order: trim → remove empty records → remove duplicates. Your header is never cleaned.

Blank means all columns. Numbers start at 1. A partial key removes the later whole record even when other fields differ.

Review the change counts before exporting. Restore turns every cleaning rule off; the raw input stays intact.

Add a file, paste some text, or try the example. Nothing is uploaded.

CSV cannot store spreadsheet column types. Confirm import settings and formula-protected values.

What exactly changes?

Trim removes edge whitespace from data fields, including tabs and edge line breaks, while preserving internal whitespace. Empty-record removal tests every field after optional trim. Duplicate removal compares all fields or selected numbered columns as strings and keeps the first whole record. The header is excluded. Ignore case compares lowercase keys; output spelling stays unchanged.

Review the impact before downloading

Counts cover the whole input; examples show at most the first ten change events with source record numbers. Trimmed-cell counts include cells in subsequently removed records. Your raw input stays intact. Restore original preview turns every rule off and rebuilds from the unchanged parsed source. Uneven nonempty records remain visible and block export; an entirely empty physical record is removable only when you choose that rule.

Choose safe duplicate keys → · Understand empty records and whitespace →

Check more than the first few records

The preview displays twenty data records per page and every parsed column. Scroll inside the table for wide files, use Next for later records, or filter a specific column. The source record number is a preview label, never an extra output column. Quoted line breaks belong to a field, so a record can span several physical lines.

Choose All result records to export the complete result, or Only records matching the filter to export all matches across all pages. The current preview page does not determine the download. The header, when enabled, is retained in either export. Zero matches disable a matching export.

What happens to my original file?

The selected file is read locally. Downloads create a new file; the original is not overwritten. Changing input or processing options disables export until you preview again. Clear input releases the tool's current data references.

Can this recover already-missing zeros or characters?

No. Use the original export. Neither encoding changes nor cleaning can infer which characters were lost without a separate source of truth.

Find a guide for your import problem →

Sources: RFC 4180: CSV format; Microsoft: duplicate records. Behaviors and limits above describe TableMender specifically.