Account

What new tool do you want?

Office & Documents

Split CSV File

Split a large CSV into smaller chunks by number of rows — each part keeps the header row — for imports with row limits.

How to use

  1. 1Paste the CSV and set the rows per part.
  2. 2Copy each part (they're separated by a marker line) into its own file.

Frequently asked questions

Is the header repeated in each part?

Yes, when 'Repeat header' is on — so every part imports on its own.

How large a file can I split?

Tens of thousands of rows work fine in the browser. For multi-gigabyte files, use a desktop tool or command line (split).

Why split a CSV?

Many tools cap imports — e.g. email platforms or CRMs that accept only a set number of rows per upload.

Row limits exist on the receiving end, not in the CSV format itself

CSV has no built-in row cap — the limit always belongs to whatever's importing it: a CRM capped at 500 rows per upload, an email platform capped at 2,000 contacts per batch. This tool exists purely to match a file to whatever ceiling the destination imposes, which is why 'rows per part' is a free-form number rather than a fixed preset.

Repeating the header is what makes each part self-contained

With 'Repeat header' on, every part is independently importable — each starts with the same column names, so a tool that doesn't understand CSV context still reads it correctly. Turn it off only when you plan to reassemble the parts into one file later and don't want the header duplicated in the middle of the data.

A worked example: 3 contacts split into chunks of 2

Paste name\nA\nB\nC with 'Rows per part' set to 2 and header repeat on. The tool produces two labeled parts separated by a marker line: 'part 1 (rows 1–2)' containing the header plus rows A and B, and 'part 2 (rows 3–3)' containing the header plus row C alone. For a real mailing list of 1,200 contacts going into a platform capped at 500 per upload, the same logic with 'Rows per part' set to 500 produces three parts — each pasteable as its own upload without any further editing.

It starts by cutting the sample into pairs of rows

The rows-per-part box opens at 2, and repeat-header is already on. With the sample of a header plus three contacts, that means the first part keeps the header and two people, and the next part repeats the header with whoever is left. The box will not go below 1. Each part is introduced with a marker that includes the part number and the row span, then a count of parts is reported beside the text. I use that marker when I am about to paste one chunk into a sheet that rejects a long import.

Turning the header off makes later parts harder to read

If I switch repeat-header off, only the first part keeps the column names and the later parts are bare rows in the same order. That is smaller, and it is also easy to mis-assign columns when I open a later part on its own. The split follows whatever delimiter was detected in the paste, so a semicolon file stays a semicolon file in every part. Changing the row count rebuilds the parts immediately. A completely empty paste stays empty rather than inventing a header. The output box is read-only, and once a split exists I can copy it or download it as a plain text file.

Related tools

More: Office & Documents