Add or Remove CSV Columns

1. 1

   Load the CSV

   Drop a .csv, .tsv, .txt or .psv file onto the page, or open the paste box and paste the text. The delimiter, quote character and encoding are detected; each can be overridden in the settings if the file is unusual.

2. 2

   Say whether the first row is a header

   Leave 'First row is a header' ticked for an ordinary export. Untick it for a file that starts straight into data, so its first row is treated as data rather than as column names.

3. 3

   Edit the column list

   Untick every column you want removed, drag the grip or use the up and down arrows to change the order, type over a name to rename that header, and press Add column to insert a new one.

4. 4

   Set what a new column contains

   Each new column takes a constant value, a blank, a row number counting from 1, or a copy of another column. The header row gets the name you typed and every data row gets the value.

5. 5

   Check the preview

   The table under the column list shows the first 50 rows of the result, updating as you edit. It is the output, not the input, so what you see is what the file will contain.

6. 6

   Apply and download

   Press Apply. Choose the output delimiter, line ending and whether to write a byte-order mark first if the file is going into Excel. Download the CSV, or copy it straight to the clipboard.

Everything about the output is one ordered list of columns. Keeping, removing, reordering, renaming and inserting are all edits to that list, and they all take effect in a single pass — there is no mode to switch between and no limit of one change per run.

• Remove any number of columns — Every column has a tick box. Untick the ones you do not want and they are gone from every row, header included.
• Add columns anywhere — Insert a new column and drag it to the position you want. It takes a constant value, a blank, a row number, or a copy of another column.
• Reorder by dragging or from the keyboard — The grip on the left of each row drags it; the up and down arrows beside it do the same thing without a pointer.
• Rename headers in place — Type over a column's name and the output's header row carries the new text. The data cells beneath it are untouched.
• A preview of the OUTPUT — The table shows the first 50 rows of the result, not of the input, and it updates as you edit — so the file cannot disagree with what you were looking at.
• Output written for its destination — Choose the delimiter, LF or CRLF line endings, whether to quote every field, and whether to write a byte-order mark for Excel.

Removing is unticking. The column disappears from the header row and from every data row, and the preview shows the remaining columns immediately.

• Several at a time — Most cleanup jobs are 'drop these six', not 'drop this one'. Untick as many as you like and press Apply once.
• Keep only a few — Press Remove all, then tick the two or three you want to keep. That is usually fewer clicks than unticking the rest.
• Keep the value, drop the header — Add a column set to Copy of another column, point it at the column you are about to remove, and rename it. The copy still resolves after the original is gone.
• Nothing is left behind — Rows in the file that are shorter than the header are padded, so the output is a rectangle with no ragged tail where a column used to be.

A new column has a name, which is written into the header row, and a value, which is written into every data row. The four value modes cover the things people otherwise leave the tool to do in a spreadsheet.

• Constant — The same text in every data row. This is the one for tagging a batch with its source, its import date or its campaign.
• Blank — An empty column with a header, ready to be filled in wherever the file is going next.
• Row number — 1, 2, 3 down the file. It counts data rows, so with a header row the first record is 1 rather than 2.
• Copy of another column — Duplicates an existing column by its position in the original file — so it keeps working after that column has been reordered, and after it has been removed.

The order of the list is the order of the output. Moving a row moves a column, and that includes moving a newly added column into the middle rather than leaving it at the end.

• Drag — Press the grip and move. The row under the pointer is highlighted so it is clear where the column will land before you let go.
• Keyboard — The up and down arrows move a row one place. A control that can only be dragged cannot be used without a pointer at all, which is why both exist.
• Rename — Type over the name. Only the header row changes; renaming does not touch a single data cell.
• Reset — Reset columns puts the list back to the file's own columns, in the file's own order, with their original names.

This is the one setting that changes what the rest of the tool means, and it is the setting most online column editors do not offer at all.

• Ticked — Row one is column names. The list is labelled with them, renaming writes to that row, and a new column's name goes there too.
• Unticked — Every row is data. The list is labelled by position, no header row is written, and renaming is switched off — a name with nowhere to go is a control that only looks like it works.
• Why it matters — A tool that always treats row one as a header consumes the first record of a headerless export as column names. The output looks entirely normal, which is what makes it dangerous.
• You can change your mind — Toggling the setting re-reads the names without re-reading the file, and column names you have edited yourself are kept.

A CSV is only simple until a field contains the delimiter. Fields are read to RFC 4180 so the awkward cases survive the round trip intact.

• Detection that parses rather than counts — Each candidate delimiter is used to parse a sample, and the one producing the most consistent row widths wins. Counting separators on the first line instead is how a semicolon export with a quoted address gets read as one column.
• Quoted fields — A quoted field may contain the delimiter, line breaks and doubled quotes. All three come through unchanged and are re-quoted on the way out.
• A quote only opens a field at the start of one — An apostrophe in the middle of a bare field is a character, not a quote. Treating every quote as a toggle is what merges two rows into one on a file full of names like O'Brien.
• Ragged rows — Rows of different lengths are a property of the file, not an error. Short rows are padded on output so the result is rectangular.
• Set it by hand — Comma, semicolon, tab and pipe, and a double-quote, single-quote or no quote character, are all available if the detection reads a file wrongly. The tool says so when it is unsure and names the runner-up.

Most 'the accents came out wrong' problems are an encoding mismatch at one end or the other, and the tool addresses both ends.

• The file's own mark wins — A byte-order mark is the file stating its encoding; a dropdown is a guess. The mark is read first and overrides the choice, and the tool says when that happened.
• Real ISO-8859-1 — In a browser, iso-8859-1 is a label for Windows-1252, so the two ordinary options are the same decoder. True Latin-1 — bytes 0x00 to 0xFF straight onto the matching code points — is provided separately, which is the only way to read a genuine Latin-1 file correctly.
• UTF-16 in, UTF-8 out — Excel's 'Unicode Text' export is UTF-16. It is read correctly and written back as UTF-8, which is what everything downstream expects.
• Writing for Excel — Tick Add a byte-order mark and Excel opens a UTF-8 CSV as UTF-8 instead of as legacy Windows text. CRLF line endings are the matching choice for Windows programs that are strict about them.

The output is written for wherever the file is going next, not just re-emitted in whatever shape it arrived.

• Delimiter — Keep the file's own, or write comma, semicolon, tab or pipe. The download's extension follows: a tab-separated file is named .tsv, a pipe-separated one .psv.
• Line endings — LF for anything Unix-shaped, CRLF for Windows programs and for strict importers.
• Quote every field — Off by default, because gratuitous quoting makes a diff unreadable. On for importers that insist on it.
• Copy instead of downloading — The result can go straight to the clipboard. The clipboard gets the text without a byte-order mark, so nothing invisible is pasted into the next window.

Columns are mapped and the file rewritten off the main thread, so the page stays responsive while a large export is processed.

A column list built once can be applied to a whole queue of exports, which is the shape of the job when a report arrives split by month or by region.

• One list, every file — The list is built from the first file and applied to the rest, with each file's own column count respected so a shorter file cannot read past its end.
• Delivered as one archive — Results come back as a single ZIP rather than as one save prompt per file, with duplicate names made unique so nothing is silently overwritten.
• Free is one file at a time — Every column feature, every value mode and every output option is available on the free plan. What Pro buys is scale and delivery.

Adding a column labels or categorises rows. Removing one drops a field before the file goes somewhere it should not. Reordering and renaming are what an importer usually needs before it will accept the file at all.

• Cleaner exports — Strip the columns the next tool does not want, and the ones you would rather not share.
• Quick enrichment — Append a constant tag, a date or a running number across every row.
• Matching an import template — Most importers want particular column names in a particular order. Renaming and dragging gets a file there without opening a spreadsheet.
• Removing personal data — Dropping an email or a phone column before sharing a file is one tick box, and it all runs on your device, not our servers.

• Column list — Tick to keep, drag or arrow to reorder, type to rename, Add column to insert. Keep all, Remove all and Reset columns act on the whole list.
• New column value — Constant, blank, row number, or a copy of another column.
• First row is a header — On for an ordinary export; off for a file that starts straight into data.
• Reading — Delimiter auto, comma, semicolon, tab or pipe. Quote auto, double, single or none. Encoding auto, UTF-8, UTF-16LE, UTF-16BE, Windows-1252 or ISO-8859-1.
• Writing — Output delimiter same as input or chosen, LF or CRLF, optional byte-order mark, optional quote-every-field.
• Delivery — Download, copy to clipboard, or a single ZIP for a batch. Custom filename with Pro.

• CSV Escape / Fix
• CSV Viewer
• CSV to JSON
• CSV Tools
• Now filter the rows

  Columns are only half the cleanup — keep the rows that match a value

• Sort by a column

  Order the file by the column you just renamed or moved to the front