Split CSV File

Every import form has a limit. Five thousand rows here, ten megabytes there, and the export you are holding has ninety thousand rows in it. I built this the week I lost an evening to a mailing platform that accepted ten thousand row chunks and offered no way to make them, while my spreadsheet app kept freezing on the file. Drop a CSV here, say how many rows you want in each part or how many parts you want, and download every piece in one zip.

Before it splits anything it tells you what you have: the row count, the column count, and the sizes of the parts you are about to create. The header row is copied onto every part, because a chunk of data with no column names is a file nobody downstream can use. Quoted fields are handled properly, including the awkward ones containing commas or line breaks. It runs in your browser and nothing is uploaded. If your data is currently a spreadsheet, the Excel to CSV converter is the step before this one.

How to use

1. Drop your CSV onto the box, or press the button and choose the file.
2. Check the row count and column count that appear, and confirm they match what you expected.
3. Choose how to split: a fixed number of rows in each file, or a fixed number of files.
4. Enter your number. If an import tool set the limit, use a value slightly under it for safety.
5. Review the list of parts and their sizes before you commit.
6. Press the button to build the parts, with the header row copied onto each one.
7. Download the zip, unzip it, and open one part to confirm the columns line up before you upload anything.

Why use our split csv file?

The header on every part is the feature people do not think about until it is missing. Split a file naively and part one has column names while parts two through nine are anonymous columns of data, which means every import after the first either fails or silently maps the wrong fields. Copying the header onto each part makes every piece a valid standalone CSV.

The part I got wrong first was the parsing, and it is the reason most quick split scripts produce broken files. My first version cut the text on commas and line breaks, which works until a row contains an address like 12 High Street, Apt 4, or a customer note with a line break in it. Those fields are wrapped in quotes precisely so the comma and the newline inside them are data rather than structure, and a naive splitter cuts through the middle of a record. This one follows the rules in RFC 4180, so quoted commas, quoted line breaks and escaped quotes stay inside their fields.

Splitting by row count and by file count are different jobs and both are here. Row count is what you want when a platform imposes a limit, since you set the number just under the cap and stop thinking about it. File count is what you want when dividing work between people, because five roughly equal files beats seven where the last one has eleven rows in it.

Nothing is uploaded, and on this tool that is not a nice extra. Think about what is inside a large CSV: customer names, email addresses, phone numbers, order histories, salaries, patient identifiers. Uploading that to a free website to perform an operation as trivial as cutting a file into pieces is a genuine data transfer, one that in many organizations would need a review and a contract. Here the file is read by your browser and split on your machine. There is no server copy because there was never a server.

You also get the numbers before you commit. Row count, column count and the size of each part are shown up front, which catches the ordinary mistakes: the export that has one column because it was actually semicolon separated, the file with four hundred rows when you expected forty thousand, or parts that will still be too big for the limit you are aiming at. If you need the data in another shape rather than in pieces, the CSV to JSON converter answers the same problem differently.

Who is this tool for?

Import limits drive most of the traffic to a tool like this. Email platforms, CRMs, e-commerce back ends and ad platforms all cap the rows or the megabytes they accept in one go. Splitting a ninety thousand row contact list into nine clean parts turns a blocked afternoon into nine uneventful uploads.

Dividing work is the second case. A team of five cleaning a product catalog wants five files, not one file and an argument about who is editing which rows. Support and sales teams split a call list the same way, and because every part carries the header the pieces merge again afterwards without anyone rebuilding columns by hand.

Then there is the file that simply will not open. Spreadsheet apps have row ceilings and older laptops have memory ceilings, and a multi hundred megabyte export hits both. Splitting it into workable pieces is often the only way to see what is inside, and it is how people prepare data for tools that expect smaller inputs, including anything heading through the JSON to CSV converter in the other direction.

Developers and analysts use it for staged testing: import one part first, confirm the mapping is right, then push the rest. A much cheaper mistake to make on two thousand rows than on ninety thousand.

Frequently asked questions