Clean a CSV without damaging its columns
Delimiters, quoted values, duplicate rows and safe exports.
VertexFiles editorial · Updated 6 September 2026
A comma is not always a column boundary
CSV fields can contain commas and line breaks inside quotes. Splitting the text on commas by hand can corrupt those records. This workspace uses a CSV parser that understands quoted fields. If automatic detection produces unexpected columns, choose the input delimiter explicitly and parse again.
Decide whether the first row is a header
The header setting controls which row is kept aside during sorting and duplicate removal. If your data has no header, turn it off. Review columns containing identifiers, postal codes and long account numbers. These are often text rather than quantities and should not be reformatted as ordinary numbers.
Clean in small steps
Trim spaces, remove exact duplicate rows and sort one column at a time. Inspect the row count after each action. Duplicate detection compares complete rows; it does not know whether two different customer records represent the same person. The source text remains available if you need to parse the original again.
Treat spreadsheet formulas as a security boundary
Text beginning with certain characters can be interpreted as a formula by spreadsheet software. CSV and XLSX exports from this workspace neutralize formula-like text. JSON is exported as arrays of rows so duplicate header names do not overwrite one another. Test the exported file in its destination application.