A CSV editor that reads the file's own dialect
There is no single CSV. A bank exports semicolons in Windows-1252, a database dumps tabs in UTF-8, a mainframe job writes pipes with a byte-order mark, and a survey tool puts a comment line on top. Intact CSV Editor works out the dialect when the file opens and shows every part of it in the Format inspector, where you can correct it.
| Separators | comma, semicolon, tab, pipe, colon, space, unit separator, or any string you type |
|---|---|
| Quoting | ", ' or none; quotes escaped by doubling ("") or by backslash (\") |
| Line endings | LF, CRLF, CR, and files that mix them |
| Encodings | 133 — every Unicode form, Windows and ISO code pages, classic Mac OS, DOS, East Asian multi-byte, and 52 EBCDIC pages — and a file that is two of them at once |
| Byte-order marks | detected on read, preserved on write |
| Header rows | detected by comparing the first row's types with the body |
| Comment lines | an optional prefix; comments are remembered and written back in place |
| Awkward files | ragged rows, stray quotes, unterminated quotes and blank lines are read rather than rejected; the cells a short row never had are hatched |
When a file is not quite right
Some files arrive damaged, and the damage is rarely the file's own fault. The app has a tool for each of the common ways:
- The wrong character set. When detection is unsure, every character set that can read the file is listed
side by side with how well its reading fits, each showing the first line with accents in it —
Łódźbeside£ódŸsettles it at a glance. - Text written twice over.
Zürichis UTF-8 that was read as Windows-1252 and saved again. Repair Text Written Twice Over puts it back toZürich, showing what it will change first. - A stray quote. One unclosed quote can swallow the rest of a file. Edit the File as Text opens the text beside the grid; mend it there and read it back as a table with ⌘↩.
- A line break in the wrong place. Join the Rows a Line Break Cut in Two puts a record back together where an unquoted line break ended it early, and Select the Short Rows finds the rest.
Saved back byte for byte
Open a CSV and save it without edits and you get the original bytes back. That is covered by automated tests
for every preset, for Windows-1252 and UTF-16 sources, for values with line breaks and doubled quotes inside them,
and for a file of a million rows. When you do edit, the values you did not touch are written back exactly as they
were: 00731 keeps its zeros, 1/2 stays a fraction, 202609100000012345 keeps
every digit.
Quoting is yours to set: as needed, always, non-numeric values, never, or only for separators — and per column or per cell on top. Quote the Columns the File Quoted notices which columns arrived quoted all the way through and writes them back the same way.
TSV, PSV, ASV and European CSV
Tab-separated and pipe-separated files are CSV files with another separator, and open the same way. So do the
semicolon files European Excel writes, with decimal commas in them — 1.234,50 is read as a number where
a number is wanted, and kept as text in the file. ASCII Separated Values use the unit separator
0x1F between values and the record separator 0x1E between records, so a comma, quote or
line break inside a value is just what it looks like; the app reads and writes them too.
Folders that always hold the same kind of file
Exports from one system tend to land in the same folder, in the same shape. Settings ▸ Rules says so once:
a folder or a pattern such as ~/Exports/DE or *.tsv, with the separator, encoding and line
endings to read them with. Detection is set aside for those files, and the window says which rule decided.
A shape you meet often can also be kept as a named format — semicolons, Windows-1252, no quotes — chosen from the menu for any file, or used for every new one.
Everything you do with a CSV
- Edit in a native grid: type over cells, fill down, drag rows and columns, paste transposed, undo everything.
- Find and filter with regular expressions, conditions and a filter on every column heading — or search and
replace across every file in a folder, with a
.bakcopy of each. - Clean up: duplicates by the columns that matter, trim, change case, split and join cells.
- Ask questions with SQL across files, summaries, statistics and lookups.
- Compare two versions of a file, value by value, and chart the columns that hold numbers.
- Repeat the same clean-up on next month's file with a recipe of the steps you took.
- Export to 38 formats — Excel, JSON, Parquet, SQL, PDF — with eleven saved presets, RFC 4180 among them.
- Open huge files: 5.2 million rows on screen in under a second — or only the part you need, or a log that is still growing.
