in reply to A bug in DBD::CSV?

It appears you're checking errstr more often than you should. Your code works just fine if you just test the results of the fetches. I'm sure that if a result came back undefined or empty you'd want to check errstr then. This sounds like doing die $! if $! which fails for the exact same reason - if you haven't just done a system operation then you don't know that $! will be false.

In preference to doing manual testing of of success/failure, it's much easier to ask DBI to throw exceptions for you: DBI->connect( 'DBI:CSV:f_dir=.', undef, undef,{ RaiseError => 1, ShowErrorStatement => 1 } ).

⠤⠤ ⠙⠊⠕⠞⠁⠇⠑⠧⠊

Replies are listed 'Best First'.
Re^2: A bug in DBD::CSV?
by EvanCarroll (Chaplain) on Apr 13, 2006 at 07:11 UTC
    The problem isn't just the error, The error is the symptom of the problem. I was altered to this because the rows below this problem were not read at all. I will supply more details tomorrow.


    Evan Carroll
    www.EvanCarroll.com

      I copied your exact code and data into files. It worked great as long as I didn't check for the failure text when there wasn't a failure.

      ⠤⠤ ⠙⠊⠕⠞⠁⠇⠑⠧⠊