csvquote

Zero-dependency Python library for RFC-4180 CSV field quoting/unquoting plus single-line split_row, with strict and lenient modes.

Install

pip install csvquote

Requires Python 3.10+. No runtime dependencies.

Quick example

from csvquote import quote, unquote, split_row

quote("hello")                     # "hello"
quote("a,b")                       # '"a,b"'
quote('he said "hi"')              # '"he said \\"\\"hi\\"\\""'
quote("line1\nline2")              # '"line1\nline2"'

unquote('"a,b"')                   # "a,b"
unquote('"he said \\"\\"hi\\"\\""') # 'he said "hi"'

split_row('a,"b,c","d""e",f')       # ["a", "b,c", "d\"e", "f"]
split_row('\\"strict\\"', strict=True)  # raises CsvQuoteError

Quality

• 123 tests, 100% line coverage
• Zero runtime dependencies
• Strict / lenient modes; clear error messages

API

quote(field: str) -> str

Quote a field per RFC 4180. Wraps in double quotes when the field contains commas, newlines, double quotes, or starts/ends with whitespace.

unquote(field: str, *, strict: bool = True) -> str

Unquote a field. In strict mode, malformed quoted strings raise CsvQuoteError. In lenient mode, returns the field as-is.

split_row(line: str, *, strict: bool = True, delimiter: str = ",") -> list[str]

Split a single CSV line into fields, handling quoted commas and escaped quotes ("").

CsvQuoteError

Subclass of ValueError.

Running tests

pip install -e ".[dev]"
pytest                           # 123 tests
pytest --cov=csvquote            # 100% line coverage

License

MIT