Re: joining records depending on special conditions - a job for gawk?

[David Resnick]

olli wrote:
> Hello
>
> for special purposes I "constructed" a TAB-delimited text file - here a
> little excerpt with fieldnames:
>
> mainno    subno   asciino(a)   asciino(b)   ...
> 000027    00      66           69
> 000027    00      66           69
> 000027    00      66           69
> 000035    00      154          100
> 000035    00      154          100
> 000035    00      154          114
> ...
>
> For every distinct combination of mainno & subno I need a count of the
> occurences of asciino(a..NF) in a special range (e.g.: count only if $X
> between 128 and 255 and not in (196,225))
>
> the result should look like
>
> mainno    subno   COUNT
> 000027    00      5
> 000035    00      19
> 000426    01      0
> 000701    03      77
> ...
>
> I don't want to use a database for this!
> I know how to solve this problem if I had mainno & subno (eg. 000027
> 00) only on one line as $1 & $2 followed by the contents of all
> asciino-fields belonging to this combination of mainno & subno.
>
> Any ideas how to get rid of this problem?
>
> Thanks in advance
>
> oliver

How about something like this?  Or did you want gawk instead
of perl for some reason?

#!/usr/local/bin/perl
use strict;
my %combos;
while (<>) {
    chomp;
    my ($mainno,$subno,$asciinoa,$asciinob) = split /\s+/, $_, 4;
    if ($asciinoa > 128 && $asciinoa < 255 &&
        !($asciinoa >= 195 && $asciinoa <= 255)) {
        $combos{"$mainno $subno"}++;
    }
}
foreach (sort keys %combos) {print "$_ $combos{$_}\n";}

-David