The quoting trap when switching delimiters

Convert a CSV file to TSV and watch most of your quote characters vanish. A field written "New York, NY" was quoted in CSV only because it contained a comma — once the delimiter becomes a tab, that comma is harmless, so the quotes come off. This is correct, and it is the main visible effect of the conversion: the output is markedly less cluttered than the input.

Here is the trap. A field that contains an actual tab character will be quoted in the TSV output, and that is where most consumers fall over. The convention for TSV in the wild is that a tab is a hard delimiter and fields simply never contain one, so a great many TSV readers implement no quoting at all. A technically-correct quoted field can therefore still break the very tool you paste it into — the reader sees the tab, splits the column, and your row silently gains a field. The same applies to embedded newlines: they are preserved correctly inside quotes, but a line-oriented reader will treat them as a row break.

The practical advice is to search your data for embedded tabs and newlines before trusting the output. If you find any, decide deliberately whether to strip them, replace them with spaces, or use a format that handles them — rather than discovering the problem downstream.

CSV parsing: what it handles and what it does not

The parser implements the useful parts of RFC 4180: quoted fields, escaped quotes written as "", and delimiters or newlines appearing inside quotes. A quote only opens a quoted section at the start of a field, so an inch mark or apostrophe in the middle of a value is treated as the ordinary character it is. Carriage returns are stripped, so Windows line endings normalise cleanly, and completely empty rows are dropped.

What it does not do is guess. There is no delimiter auto-detection and no encoding detection — the input is assumed to be comma-separated UTF-8 text. Paste something with no commas at all and you get No rows found. Make sure your CSV has comma separators. Note that a genuinely single-column CSV converts perfectly well; that error means nothing parsed as a row, not that you supplied too few columns.

Output rows are joined with newlines and separated by tab characters. Save it with a .tsv extension if your downstream tool expects one.

When to reach for related tools

If you want something readable rather than machine-parseable, the CSV to Markdown Table Converter renders the same data as a GitHub-flavoured table. Going the other way, TSV to CSV handles spreadsheet clipboard pastes, which arrive tab-separated.