Practical guide
Clean, Deduplicate and Sort CSV Tables
By MultiZest · Updated
How to use this tool
- Paste or load CSV up to 500 KB with a unique header row, then choose the delimiter.
- Choose cleanup rules, an optional duplicate key and an optional sort column.
- Clean the table, check removed-row counts and preview, then download CSV or TSV.
Worked example
Rows with customer code 00123 keep that text value. If you deduplicate using the code column, the first row for 00123 remains and later rows with the same key are removed, even if another column differs.
Cleaning rules change your data deliberately
Trim affects headers and values. Empty-row removal ignores rows where every value is blank. Deduplication runs after trimming and keeps the first occurrence. Matching is case-sensitive. A single-column key also treats empty values in that column as equal.
Quoted fields, commas inside values and line breaks are parsed as CSV. Inconsistent row widths and empty or duplicate headers are rejected. The table is limited to 200,000 cells and the parser supports up to 20,000 data rows.
Sorting and spreadsheet export
Text sorting uses English collation. Numeric sorting requires a valid finite number in every selected cell; values remain strings in the output, so sorting does not remove leading zeros. Use text sorting for large identifiers.
Formula protection is enabled by default and prefixes risky spreadsheet cells with an apostrophe. The preview shows the cleaned values before this export protection. Turn it off only when exact unmodified text is required and appropriate for the receiving workflow.
Decide what a duplicate row means for your data
An accidental duplicate export row and two legitimate purchases with the same visible values are different situations. Before removing duplicates, decide whether exact whole-row equality is the rule you actually need. Two transactions can share a date and amount while representing separate events.
Use a source copy and test the cleaning settings on a small sample. Check rows with blank fields, padded text and different capitalization. Trimming spaces can resolve accidental padding, but it can also change a value whose spaces are meaningful. Read the preview before accepting the whole-table transformation.
Check the header, row count and sort order
Compare the number of data rows before and after cleaning and investigate a larger-than-expected reduction. Confirm that the header remains aligned with each column. A neat-looking table is not sufficient if a row lost a field or a different record was removed as a duplicate.
Sorting by a text column may differ from numerical or date sorting. Values such as 2 and 10 can appear in an order that makes sense as strings and surprises someone expecting numbers. Review the actual order and formula protection before opening the output in a spreadsheet or importing it into another system.
Questions and troubleshooting
Which duplicate row is kept?
The first matching row in the original order is kept. Sorting runs after deduplication.
Can I use tab-separated data?
Yes. Select Tab (TSV). The downloaded output uses the same chosen delimiter.
Should I remove duplicate rows from financial data automatically?
Only after defining the correct duplicate rule and preserving the source. Exact matching rows may be legitimate separate records. Review identifiers and expected totals in the receiving workflow.
Continue with a related tool
Convert CSV and JSON with quoted-field support, custom delimiters, preserved text values and spreadsheet formula protection.
Remove Duplicate Lines →Remove duplicate lines from lists or text blocks with custom sorting.
JSON Compare →Compare JSON structure by path, ignoring object key order while preserving array order. Review added, removed and changed values.
Sources and further reading
Found an error or an unclear instruction? Report it. Read our editorial policy and privacy policy.