Compare CSV files

Comparing spreadsheet exports

CSV is the format almost every system will export to, which makes it the natural common ground when you need to know what changed between two snapshots of the same data. A customer list from last month against this month, two versions of a product catalogue, an inventory count before and after a stock take.

Because each CSV record occupies one line, line-level comparison maps cleanly onto rows. A changed row shows up as a modification with the old values on the left and the new ones on the right. A row that only exists in one file shows as an addition or a deletion.

Sort both files first

This is the single most useful piece of advice for CSV comparison. A diff aligns lines in order, so if the two exports list the same records in a different sequence, almost every row will be reported as changed even though the underlying data is identical.

Sort both files by the same key column — usually an ID — before comparing. In a spreadsheet that is one sort operation; on the command line it is one sort call. Do that and the comparison goes from unreadable to obvious. Keep the header row in both files so the columns stay aligned.

Finding which field changed

Line granularity tells you a row changed. To find out which column, switch to word or character granularity — the highlighting then narrows to the specific value that differs rather than the whole row.

This is especially useful for wide exports with twenty or thirty columns, where visually scanning a modified row for the one altered cell is genuinely slow. At character granularity a changed postcode or a flipped digit in an amount becomes immediately visible.

Why local processing matters for data exports

CSV exports are, by their nature, the files most likely to contain personal information — names, email addresses, phone numbers, order histories, salaries. Under GDPR and similar regimes, sending that data to a third-party service is a processing activity you would need a basis for.

Here there is no transfer to consider. The file is read by your browser and compared on your own machine. You can verify it: open developer tools, watch the Network tab, and run a comparison. Nothing is sent.

Handling large exports

Exports of tens of thousands of rows work, though very large files take a moment because the processing happens on your device. If a file exceeds the 5 MB upload cap, split it or filter to the columns you actually care about first. There is no monthly quota — compare as often as you need.