James Edward Gray II wrote:
>
> On Saturday, October 5, 2002, at 06:02 PM, Paul Van Dalen wrote:
> >
> > Given this input:
> >
> > 222 xxx JJJJJ
> > 222 www DD
> > 222 cc ffFFF
> > 909 eee EEEE
> > 909 FFF kkkk
> > 909 jjjjj KKKK
> > 888 JJJ HHHH
> > 888 HHH JJJJ
> >
> > I'd like to split the input, group it by the first column, and, for that
> > group, print out a concatenated second column. The output should look
> > thusly:
> >
> > 222 xxx;www;cc
> > 909 eee;FFF;jjjjj
> > 888 JJJ;HHH
>
> perl -pe 's/^(\d+\s+)([a-zA-Z]+)\s+([a-zA-Z]+)$/$1$2;$3/' infile >outfile
$ perl -pe 's/^(\d+\s+)([a-zA-Z]+)\s+([a-zA-Z]+)$/$1$2;$3/' infile
222 xxx;JJJJJ
222 www;DD
222 cc;ffFFF
909 eee;EEEE
909 FFF;kkkk
909 jjjjj;KKKK
888 JJJ;HHHH
888 HHH;JJJJ
This _will_ work: :-)
$ perl -ane'
BEGIN{$prev = ""} END{print "@$_\n" for @data}
if ($prev eq $F[0]) {$data[-1][1] .= ";$F[1]"}
else {$prev = $F[0], push @data, [@F[0,1]]}' infile
222 xxx;www;cc
909 eee;FFF;jjjjj
888 JJJ;HHH
And if you don't care what order the records are output.
$ perl -lane'push@{$d{$F[0]}},$F[1]}{print"$_ ",join";",@{$d{$_}}for keys%d' infile
222 xxx;www;cc
909 eee;FFF;jjjjj
888 JJJ;HHH
John
--
use Perl;
program
fulfillment
--
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
