Delete CSV Columns

RFC 4180 is the de-facto CSV standard. It specifies: (a) fields can be enclosed in double quotes; (b) embedded delimiters/newlines are allowed inside quotes; (c) embedded quotes are escaped by doubling (""); (d) quote-enclosed fields can span multiple physical lines. Naive line/comma splitting fails all of these. Real-world CSVs from Excel, Google Sheets, government data portals, and pandas all use RFC 4180 conventions.

1-based (default): “column 3” means the third column. Matches how spreadsheets and humans count. Use this unless you have a programming reason not to. 0-based: “column 0” means the first column. Matches array indexing in most programming languages. Use for compatibility with scripts. By name: write header text exactly as it appears (case-sensitive). Works only when First-row-is-header is on. Most reliable against column reordering.

When First-row-is-header is on AND CSV has data, the parsed headers appear as clickable buttons under the options. Click any chip to add (or remove) that column from the selection. The selection field updates with the correct token for the current mode (1-based number, 0-based number, or name).

Inverts the selection logic. Without invert: “delete columns 1, 3, 5”. With invert: “delete EVERYTHING EXCEPT columns 1, 3, 5” (i.e., keep only those). Useful for the common “I just want these specific columns” workflow without listing all the columns you don’t want.

The serializer quotes a field only when necessary: if it contains the delimiter, a quote character, a newline, or a carriage return. Inner quotes are doubled (" → ""). This matches Excel’s behavior and round-trips through any RFC 4180 parser. Fields without special characters are emitted unquoted (compact output).

Handled. Each row is processed independently. If row 5 has only 3 fields but the header has 5, the row passes through with its 3 fields minus whichever match the deleted indices. Output may have varying row lengths (preserves input irregularity). Use a CSV validator/normalizer if you need uniform widths.

Pick “Tab” in the delimiter selector. Drag-drop a .tsv file and the auto-detector picks tab if present. All RFC 4180 rules still apply but with tab instead of comma. Note: very few exporters quote-and-escape in TSV (it’s typically uncommon to have tabs inside cell values), but the tool handles it if you do.

Pick “Semicolon” in the delimiter selector. Auto-detection picks semicolon if your sample contains ; but no , (heuristic – works for most cases). European Excel exports use semicolon when the locale’s decimal separator is a comma (so 1,5 isn’t ambiguous with field separation). Same RFC 4180 rules apply.

Yes. All parsing and filtering happens in your browser using vanilla JavaScript. Your CSV file/text never leaves your device. The download is generated in-memory.