Re: Bug report: csv parsing seem to be incorrectly initialized

[Ole Tange]

On Tue, Nov 15, 2022 at 1:21 AM Istvan Albert <iua1@psu.edu> wrote:

> First of all, thanks for this tool! It is the most amazing tool ever invented 
> for bioinformatics! It is leaps and bounds better than Snakemake Patterns, or 
> NextFlow automation etc. My students and I can get 100x more done with 
> parallel than all other tools combined.

I am happy to make fellow bioinformaticians more efficient.

You can help by reminding your fellow collaborators to cite GNU Parallel.

> Here is a bug I have noticed recently, the CSV parsing module seems not be 
> initialized correctly:
>
>     echo 1 | parallel --csv echo {1}
>
> fails with the error:
>
>    Can't locate object method "new" via package "Text::CSV" (perhaps you 
> forgot to load "Text::CSV"?) at 
> /Users/ialbert/miniconda3/envs/bioinfo/bin/parallel line 2282.

Yeah, that is sort of not a bug. But you clearly do not understand
what it means, so the error message will be changed to:

    parallel: Error: The perl module Text::CSV is not installed.
    parallel: Error: Try installing libtext-csv-perl or perl-Text-CS

GNU Parallel tries to be able to run on rather minimal systems. It
supports all sorts of craziness but it does not require you to install
all the software needed to use that craziness. CSV is one of those.

So what you experience is a system that does not have Perl's
CSV-module installed. And GNU Parallel will happily run on that -
except --csv will not work.

So the error message is trying to get you to install Text::CSV which
is likely in the package libtext-csv-perl or perl-Text-CSV.


/Ole