Data Format
Gold standard and classification data is read from CSV files. The particular CSV format variant can be controlled by the
set command.
With the default format, commas are used to separate values. If a single value includes commas, it must be wrapped with double quotes.
A value starting with a double quote must end with one, and double quotes within the value itself must be escaped by using two consecutive double quotes.
If a value does not start with a double quote, double quotes may occur freely within the value itself.
Examples of legal lines using the default CSV format:
1,the quick brown fox (resulting string: the quick brown fox)
2,"the quick brown fox" (resulting string: the quick brown fox)
3,"the quick, brown, fox" (resulting string: the quick, brown, fox)
4,the quick "brown" fox (resulting string: the quick "brown" fox)
5,"the quick, ""brown"", fox" (resulting string: the quick, "brown", fox)
6, "the quick, "brown", fox" (resulting string: "the quick, "brown", fox")
6, "the quick brown fox (resulting string: "the quick brown fox)
Examples of illegal lines using the default CSV format:
1,the quick, brown, fox (not strictly illegal, but will be read as multiple sub-strings)
2,"the quick, "brown", fox" (quotes part of the value must be escaped)
3,"the quick brown fox (value starting with a quote must end with a closing quote)