CSV Column Extractor

A focused cut from a wide CSV without opening a spreadsheet. Parses RFC 4180–style CSV (quoted fields with embedded delimiters and escaped quotes), with selectable tab/semicolon/pipe delimiters. Pick columns by header name or index, optionally dedupe and sort, and emit them in the format that matches the next step — newline-per-row for piping into another tool, or a (...) IN-clause to drop into a SQL editor.

Proper quote handling is the reason to use a parser rather than splitting on commas, and it is exactly where a quick script goes wrong. A quoted field may contain the delimiter, a newline, or a doubled quote standing for a literal one, so a line in the file is not necessarily a row in the data. Splitting on commas turns one address field into three columns and silently shifts every column after it, which produces output that looks plausible and is wrong.

Selecting by header name is safer than by index whenever the source might change, since an inserted column shifts every index after it but leaves the names alone. Names have to match the header exactly, which makes a trailing space or an invisible byte-order mark at the start of the file a common reason a column appears missing.

The SQL output is a convenience for pasting into a query, not a way to build queries from untrusted input. It quotes and escapes the values it emits, but a list assembled by string concatenation is the wrong shape for anything that reaches a database on a user's behalf; use bound parameters there. Long lists also run into per-statement limits in most engines, which is worth knowing before pasting fifty thousand identifiers.

Everything runs in the browser on text you paste, so the practical ceiling is your tab's memory rather than a file size limit. Very large exports are better handled by a streaming tool.