Re: Use of --colsep without --header for multiple input files

[Ole Tange]

On Sat, Jul 1, 2017 at 1:56 AM, Coby Viner <cviner@cs.toronto.edu> wrote:

> Is it possible to use --colsep with multiple input files, without having to
> manually prefix each files' header by a unique identifier for use in
> conjunction with --header?

Yes.

> For example, it would be nice if with multiple files they could be
> automatically labelled with a–z, with numerals still referring to their
> columns (the reverse strikes me as better, but that would be
> counter-intuitive given the present implementation).

The current implementation numbers the total amount of columns. In
your case {1}, {2}, {3} and {4}.

Nice thing about this is that negative number and all --rpl strings
starting with '{' work, too. Here we lowercase column 3 (second to
last = -2):

$ parallel --rpl '{lowercase} $_=lc($_)' --colsep '\t' "echo f1={1}
f2={2} g1={-2 lowercase} g2={4}" :::: <(perl -e 'printf
"f1\tf2\nA\tB\nC\tD\n"') :::: <(perl -e 'printf
"g1\tg2\nE\tF\nG\tH\n"')

> I would ideally like to be able to do the following:
>
> parallel --colsep "echo f1={a1} f2={a2} g1={b1} g2={b2}" :::: <(perl -e
> 'printf "f1\tf2\nA\tB\nC\tD\n"') :::: <(perl -e 'printf
> "g1\tg2\nE\tF\nG\tH\n"')
>
> Currently, the use of --colsep without --header for multiple files does not
> behave as I would have expected:
>
> parallel --colsep '\t' "echo f1={1} f2={2} g1={1} g2={2}" :::: <(perl -e
> 'printf "f1\tf2\nA\tB\nC\tD\n"') :::: <(perl -e 'printf
> "g1\tg2\nE\tF\nG\tH\n"')
> f1=f1 f2=f2 g1=f1 g2=f2
> f1=f1 f2=f2 g1=f1 g2=f2
> f1=f1 f2=f2 g1=f1 g2=f2
> f1=A f2=B g1=A g2=B
> f1=A f2=B g1=A g2=B
> f1=A f2=B g1=A g2=B
> f1=C f2=D g1=C g2=D
> f1=C f2=D g1=C g2=D
> f1=C f2=D g1=C g2=D

Try using {3} and {4}:

parallel --colsep '\t' "echo f1={1} f2={2} g1={3} g2={4}" :::: <(perl
-e 'printf "f1\tf2\nA\tB\nC\tD\n"') :::: <(perl -e 'printf
"g1\tg2\nE\tF\nG\tH\n"')

I have updated the man page:

       --colsep regexp
      -C regexp
               Column separator. The input will be treated as a table
               with regexp separating the columns. The n'th column can
               be access using {n} or {n.}. E.g. {3} is the 3rd column.

               If there are more input sources, each input source will
               be separated:

                 parallel --colsep ',' echo {4} {3} {2} {1} ::: "A,B" ::: "c,d"

               --colsep implies --trim rl, which can be overridden with
               --trim n.

               regexp is a Perl Regular Expression:
               http://perldoc.perl.org/perlre.html

/Ole