Hacker News Reader

Comment by gav | Hacker News Reader

Note that it fails to parse CSV like:

    "x",""x"",,"x
    x","x"
    "y",,,,123

View on HN

The previous commit (probably) failed at that. However, with the latest commit, parsing that malformatted CSV returns:

    [
      ["x","x\"","","x"],
      ["x","x"],
      ["y","","","",123]
    ]

I can't see anything wrong with it. Am I missing something?

It should parse as:

    [
      ["x","\"x\"","","x\nx","x"],
      ["y","","","",123]
    ]

It's valid CSV, inasmuch as it's accepted and generated by a lot of tools and RFC 4180 ABNF:

    escaped = DQUOTE *(TEXTDATA / COMMA / CR / LF / 2DQUOTE) DQUOTE

I might be mistaken, but the second field in the first record looks invalid in that CSV. You'd need """x""" for that to work (quoted field and escaped quotes inside).

You're right. Removing the claim of compliance until I've devised a solution.

(Are you talking about Papa, seeing that your comment is a response to mine?)

That CSV text is malformed. Arguably, the parsing succeeded, but the expected errors were generated and reported.

Reply on news.ycombinator.com