CSV CLEANER + EXCEL CONVERTER
Remove duplicates, choose the columns you need, and rename or reorder them for your next import. Download text-safe Excel, CSV or TSV without uploading your data.
Local processing. No account.
Free CSV, TSV and Excel downloads.
01 / INPUT & SETTINGS
UTF-8 and UTF-16 BOM detection. Other encodings can be selected below.
Quoted commas, double quotes and multiline cells are supported.
Run once to see your columns. With none selected, every column must match. Matching is case-sensitive, after any selected trimming.
Your columns will appear here.
The recommended option removes exact duplicate rows but preserves different values sharing a key. Keep-first can discard newer or more complete data; it does not merge cells. A missing or whitespace-only key part always keeps that record. Removing blank rows is a separate choice.
Add a file or paste data to get started. Nothing is changed by default.
02 / REVIEW & DOWNLOAD
Your rows will appear here, with a clear count of what changed. IDs, dates and amounts stay as text.
Preview your full source first. Then keep only the fields the destination needs. Reordering or omitting columns never changes duplicate matching: it still uses the full source table.
This selects exported fields; it is not a redaction service. The source workspace and original-value removed-row log can still contain omitted cells. The optional column map includes header labels but no cell values.
CHECK WHAT CHANGED
Original cell values are shown before trimming. Highlighted columns differ within a key group. Record numbers include the header, not physical lines inside quoted cells.
Review all pages in a group. “Use record” keeps that whole original record and removes the rest; it never merges fields. Changing source data or cleanup settings resets choices. Export column changes preserve them. Exact repeats remain removed after resetting choices.
The JSON log retains original cell text, IDs, row numbers and removal reasons. It is a review log, not a spreadsheet. It does not replace your original file.
Excel stores every cell as text: 00123 stays 00123, long IDs stay intact, and formula-like values never run.
Excel is prepared only when you request it. CSV and TSV are ready with the preview.
A WORKED EXAMPLE / SYNTHETIC DATA
A contact export can contain a repeated record, an empty row and IDs that a spreadsheet mistakes for numbers. The built-in sample includes all three. Here is what changes—and what must not.
Two identical Ada records, one blank row, and three other contacts. Ada’s name has a space on each side.
00123 | Ada
00456 | Lin
00123 | Ada
|
00007 | Sam
12345678901234567890 | JoThe first Ada record stays. One duplicate and one blank row are removed; names are trimmed. Every ID keeps its digits.
00123 | Ada
00456 | Lin
00007 | Sam
12345678901234567890 | JoOnly the ID and name columns are shown here. The sample’s email and notes columns are retained too, including a quoted comma and a multiline note.
Sam’s note =1+1 stays literal text in Excel; it does not become 2. CSV/TSV with the formula guard on adds an apostrophe instead. If you match by ID alone, different notes are preserved by default and flagged for review. Choose keep-first only when you intend to discard later conflicting records.
A DIFFERENT ID IS EASY. A CONFLICT IS NOT.
Our second synthetic example contains four records: two different entries for 00123, one exact repeat and another customer. Matching by ID keeps both different entries and removes only the repeat: four rows become three. Keep-first would leave two rows by discarding the alternative email and amount too.
Loads synthetic data, selects the ID column and turns on safe duplicate removal. It does not choose which email or amount is correct. Review the original records, then choose “Use record 3” to keep the newer entry only. “Keep distinct records in this group” restores both different entries, while still removing the exact repeat.
A FILE THAT FITS THE NEXT SYSTEM
Our synthetic import example has three customers and four columns. Open the column editor, uncheck internal_note, move customer_id earlier, and rename it Customer ID. Apply columns to create all three downloads with the same three fields. IDs 00123, 00007 and 12345678901234567890 stay text in Excel.
Changed schemas require unique, nonblank selected headers. New names are limited to 200 characters without outer spaces, line breaks or control characters. The unchanged default retains original headers, even blank or repeated ones. Headerless input stays headerless. Changing an input or header setting resets the column plan; editing a plan immediately locks old downloads.
Why these choices matter: Microsoft: choose or remove columns · Microsoft: keeping leading zeros and large numbers. CSV has no cell types; choose Excel when an ID must remain text.
SMALL FIXES. NO SURPRISES.
Trimming, blank-row removal and duplicate removal are off until you choose them. Exact duplicate matching keeps the first row. Your source file is never overwritten.
CSV files have no cell types. Spreadsheet apps can guess wrong about product codes, dates and long numbers. The Excel download writes text cells explicitly. Convert selected columns to numbers in Excel when you need calculations.
Up to 5 MiB, 50,000 data rows, 100 columns and 500,000 cells. Each cell may contain up to 32,767 characters. Outputs are limited to 40 MiB each. Invalid quotes, mismatched columns and unreadable text are explained—not silently fixed.
When the formula guard is on, values that could be interpreted as spreadsheet formulas get a leading apostrophe. This changes those text exports and reduces accidental formula execution. Behavior varies by spreadsheet app, so Excel (.xlsx) is the recommended option: it preserves the original value as a text cell with no formula. CSV and TSV do not guarantee preservation of leading zeros when opened in a spreadsheet.
No. Parsing, cleaning and file generation happen in this browser tab. File contents and names are not sent to FastTool. Your source and results are not saved between visits. Clear removes them from the workspace; already downloaded files remain on your device.