What it does
A delimited column extractor turns a table-shaped block of text into a narrower table. You name the fields you want, and the tool returns one row per input row with just those fields. It is the browser equivalent of selecting a few columns in a spreadsheet and pasting them somewhere else — without opening a spreadsheet.
The input can be comma-separated, tab-separated, semicolon-separated, pipe-separated, or split by a custom string. The output is plain text with a separator you choose. That makes it useful before a CSV import, after copying a table out of a web page, or when a log line has more fields than you need.
Delimiters and quoting
Automatic detection looks at the first few rows and chooses the most frequent candidate among tab, comma, semicolon, and pipe. It is a convenience, not a promise. If a quoted address contains several commas, the detector may still choose comma — which is correct — but a text field full of pipes can tilt the count. When the shape matters, choose the delimiter explicitly.
With quote handling on, a field wrapped in double quotes may contain the delimiter, and two consecutive double quotes inside it become one literal quote. This covers the common CSV convention. It does not implement every dialect: a backslash escape, a single-quote quote character, or a multi-character custom delimiter follows the literal path instead. If the output looks shifted, check the source for an unusual quoting style before blaming the columns.
On the output side, a selected field that contains the output separator is wrapped in double quotes, and any internal quote is doubled. That keeps a comma inside a name from looking like an extra column. Turn this off if you want a literal join for display rather than a CSV-compatible result.
Selecting columns
Fields are numbered from 1. The selector accepts a comma- or space-separated list, and the order of the list becomes the order of the output. For example, 2,1 swaps the first two columns, and 1,3 drops the middle one. Repeated numbers are ignored.
Negative numbers count from the right. -1 is the last field, -2 the one before it. This is handy for a log where the useful value is always at the end but the number of leading fields changes. A missing field becomes an empty cell and is counted in the statistics bar. If every selected cell in a row is empty and “Remove rows with no selected content” is on, the row is dropped.
Examples
Before
Name,Team,City
"Ada, L.",Research,London
Grace,Engineering,New York
Alan,Research,ManchesterAfter
Name,City
"Ada, L.",London
Grace,New York
Alan,ManchesterThe quoted name keeps its internal comma. The team column disappears, and the remaining fields are rejoined with a comma. The header row is treated like any other row because “skip header” is off; that is why it still appears in the output.
Before
Ada,Research,London
Grace,Engineering,New YorkAfter
London Research Ada
New York Engineering GraceMoving fields is the same operation as selecting them. The tool does not rename columns or add a new header. If the destination needs a header, add it there or keep the original header row and reorder it with the same selector.
Limits worth knowing
- Plain text only. It reads pasted text, not an .xlsx workbook, a Google Sheet, or a database connection. Copy the cells as delimited text first.
- Missing fields are not invented. A short row gets empty cells for the columns it does not have. A long row keeps its extra fields in the input but ignores them unless you select them.
- Quoting is a convention, not a type system. A field that looks like a number stays a string, leading zeros are preserved, and a date is not reformatted. That is usually what you want before an import.
- Multi-character delimiters bypass quote handling. If your data uses both a long separator and quoted fields, normalize the quoting first or choose a single-character delimiter.
- Header handling is a single checkbox. It skips the first row and does not try to match column names. If the header is on row two, extract it separately.
Page updated