> P.S. As Simon noted, it seems to be defined here: > <https://tools.ietf.org/html/rfc4180>
> I was probably referring to the first sentence of chapter 2: "While there are > various specifications and implementations for the CSV format" OK for that document existing and all, but just a quotes: "Each field _may or may not_ be enclosed in double quotes" So if I want to represent "" I can use either "" or """""". Noting that if I cannot tell the parser if I am using quotes or not """""" becomes either """""" or "". And if I need to use a crlf in there, quotes are mandatory. So it may or may not be enclosed but must be in some scenarios. Fantastic. This kind of BS (at least to me) reduces the relevance of a "definition". It also may or may not have headers, but that's a smaller issue.