Expand description
What a CSV file’s punctuation is, and working it out from the bytes.
A CSV file does not say how it is written. The delimiter, the quote, the escape and whether the
first line is a header are all conventions, and a reader that demands to be told them is a
reader every loader script has to be rewritten for. DuckDB sniffs, so this sniffs, and every
rule here was read off duckdb v1.4.1’s sniff_csv rather than reasoned about.
The candidates are DuckDB’s: comma, pipe, semicolon and tab for the delimiter, and the double
quote for the quote and the escape. A file with no quote character in it is reported as having
no quote at all rather than as having the default one, which is visible in sniff_csv and is
reproduced because the same field is printed back in an error message.
Structs§
- Dialect
- How a CSV file is punctuated.
- Given
- What the caller said about how the file is written, where the sniffer would otherwise decide.
Constants§
- DELIMITERS
- The delimiters tried, in the order they are tried.