CSV / TSV
Swap the entire grid, including the first row and first column. Choose how to display headings without removing values.
Each side: 1 MiB UTF-8, 1,000 rows, 1,000 columns; padded area ≤100,000 cells.
Cells are shown as literal strings. Heading view preserves the first row in exports.
No result yet.
Your table is processed in this browser; no input data is uploaded.
Paste a table to begin.
A monthly report may place dates across the top while the next chart expects them down the side. CSV transpose changes the orientation of the existing grid: the value at row 2, column 3 moves to row 3, column 2.
CSV rows to columns and CSV columns to rows are the same operation. This page keeps strings intact, so an identifier such as 007 stays 007 even when its position changes. It does not group repeated names or calculate a pivot table.
CSV transpose includes the whole first row. Its cells become the first column, while the original first column becomes the top row. Mark that output row as headings in the table preview when it contains labels; the exported values remain identical in both heading modes.
A short record must not pull later values into the wrong column. CSV transpose pads missing trailing positions with empty strings up to the widest record and reports how many were added. Use strict mode instead when unequal row lengths indicate a broken export that needs correction.
Input and output delimiters are independent. Select tab for pasted spreadsheet cells, or semicolon for a regional export. The parser does not guess: a wrong separator can make the input look like a single column. Inspect the reported input dimensions before using the result.
Padding adds only missing trailing cells; an existing empty string is left as it is. Strict mode rejects any record whose width differs from the widest record. A blank physical line is a one-cell record, while a newline inside quotes belongs to that cell.
An input of 4 rows by 3 columns becomes 3 rows by 4 columns, including any headings. The preview grows by 50 rows or 20 columns, but exports always include the full result. Automatic processing waits 300 ms after edits; turn it off to run with the button or Ctrl/Cmd + Enter. Edits expire exports until processing succeeds again.
Treat every output cell as data, or display the original first column as the new header row. Neither mode skips, renames or duplicates cells. Blank and repeated headings are allowed. CSV and TSV carry no header metadata, so the exported text is the same for both views.
Quoted delimiters, doubled double quotes and embedded CRLF or LF stay inside their original strings. Leading zeros, surrounding spaces and Unicode remain unchanged. Quote-all output is optional; fields that need quoting and empty fields are always quoted.
Use comma, semicolon, pipe or tab independently on each side. To transpose TSV, choose tab as the input delimiter; tab output downloads as csv-transposed.tsv. Other separators download as csv-transposed.csv. UTF-8 output uses CRLF between records, without a BOM or final record terminator.
Copy the result back into the input and choose its delimiter to transpose again. With the same delimiter and quoting policy, the rectangular cell matrix returns to its original orientation. Padding remains as real empty cells; original quote choices and record terminators are not restored.
Paste delimited text and select its separator. The entire matrix is transposed locally. Check the input and output dimensions and the padding count, then copy or download the current result. No spreadsheet formulas or installed application are required.
It becomes the first output column, including its top-left cell. The original first column becomes the first output row. Choosing the heading view only marks that output row as table headings; it does not remove a record or add metadata to CSV.
Yes, in the default padding mode. For a width of three, a one-cell record gains two empty trailing cells before the axes are swapped. Choose strict mode to receive an error instead. Existing empty cells and missing positions become indistinguishable once the padded file has been exported.
Both directions allow up to 1,000 records and 1,000 columns, with a maximum rectangular area of 100,000 cells and 1 MiB each for input and output. A 1,000 by 100 table fits the shape limit; a 1,000 by 101 table does not. A large individual string or added quotes can still exceed the byte limit.
Yes. Select comma, tab, semicolon or pipe for each side independently. Tab output produces TSV. Quoting follows the same rules for every delimiter, so a tab inside a quoted value remains part of that value instead of creating another column.
Found a case with the wrong cell positions?
Check one known value, the header positions and any added empty cells before exporting the complete result.