Extract specific columns from a CSV

Type the column names you want to keep, comma-separated, and click Extract columns. Output columns come out in the order you listed them, which doubles as a way to reorder a file. If a name does not exist you are told which one, rather than getting a silently empty column.

The names must match exactly

Matching is exact, including case and spaces. Email will not match a column headed email, and a header written customer name needs that space. Scan the file first and copy the names from the preview table — most failures here are an invisible trailing space in the original header.

Stripping personal data before you share a file

This is the most valuable use. Analytics exports routinely carry names, emails, phone numbers and addresses alongside the two or three columns anyone actually needs. Sending the whole file to an agency, a contractor or a colleague distributes all of it.

Keeping only the needed columns is real data minimisation, and it happens entirely on your machine — the original never leaves it, and neither does the reduced copy until you send it. Under GDPR, not sharing the data at all is considerably simpler than justifying why you did.

It also makes everything downstream faster

A 60-column export cut to six is roughly a tenth of the size. It imports faster, opens in a spreadsheet without complaint, and if you are converting to JSON — where every key is repeated on every record — the saving is larger still.

Related: Remove duplicate rows · Split a large CSV

Common questions

What if a column name is not found? You get an error naming the missing columns and nothing is exported, so you never end up with a silently empty column.

Can I reorder columns? Yes. Output follows the order you type the names in, so listing all of them in a new order reorders the file.

Can I drop columns instead of keeping them? Not directly — list the ones to keep. On a wide file, copy the headers from the preview and delete the ones you do not want.