Convert CSV Rows to Columns

Every input row becomes a column in the output – equivalently, every input column becomes an output row. If your source has labels running down the left and observations across the top (wide format), transposing flips it so labels are across the top and observations run down (long format). Shape N×M becomes M×N.

Mathematically the operation is identical – matrix transpose is its own inverse. Two pages exist because users search with both phrasings. Whichever you open, the mechanics are the same. Feed either tool’s output back in to restore the original layout.

Yes – that’s the default parser. It recognises double-quoted fields, allows embedded delimiters and newlines inside quotes, and decodes "" as a literal double quote. Output is auto-quoted when any cell contains the delimiter, a newline, or a double quote.

The tool finds the longest row and pads every shorter row with empty strings up to that width before transposing. The stats line reports how many empty cells were added, so jagged input doesn’t silently corrupt the grid.

Yes. An input of applenbananancherry (three rows, one column) becomes apple,banana,cherry (one row, three columns). Same applies for single-row inputs in the opposite direction.

It’s free with no account. And no – nothing is uploaded. All parsing, transposing, and serialisation happens locally in your browser with plain JavaScript. You can disconnect from the network after the page loads and the tool still works.

Up to around 100,000 cells (e.g., 1,000 rows × 100 columns) completes well under a second. Beyond that the bottleneck is usually the output textarea repainting, not the parse. For truly massive datasets, stream-splitting tools like DuckDB or awk are a better fit.

Yes. Any mixture of rn (Windows) and n (Unix) is normalised to plain n before parsing, so the output uses consistent line endings. A UTF-8 BOM at the start of the input is also stripped automatically.

In RFC 4180 mode, no – a transpose-then-transpose round trip is lossless for standards-compliant CSV. In Simple mode, round-trips corrupt any row that contains quoted delimiters (because the first pass splits them wrongly). That’s why RFC 4180 is the default.

Common ones: flipping a timeline CSV (dates as column headers, metrics as rows) into a “long” format that charting libraries want; restructuring exports from old survey tools; rotating a vertical key/value list into a proper table; or preparing data for a pivot operation that expects the axes swapped.