http://www.perlmonks.org?node_id=1098323

randome has asked for the wisdom of the Perl Monks concerning the following question:

Hello, I have a huge file with over 700,00 entries with 18 columns. One column is in the format 
--+-+
---++
?????
-????

[download]
Now i have another list which corresponds to this code in a particular order 
A-1
B-7
C-11
D-3
E-100

[download]
Now I need to match the expression above to the pattern, where +/- corresponds to A-E and ? corresponds to 0 and as an output I need to get a number. So the output should be 
--+-+ = 122  (because 1+7+11+3+100)
---++= 122
?????= 0
-???+= 101

[download]
This is just a sample, the numbers are much bigger and the pattern runs upto 24 times. Is there any way I can do it with perl scripting ? Thank you

Replies are listed 'Best First'.
Re: decoding pattern expression
by kennethk (Abbot) on Aug 22, 2014 at 14:25 UTC
    Is there any way I can do it with bash scripting ?
    If you want bash scripting, you've come to the wrong site. If you'd like some help doing this with Perl scripting, I'd probably use indices 0-4 instead of A-E, make an array (my @values = (1, 7, 11, 3, 100)) and then split your coded strings: 
    my @terms = split //, $input;
my $sum = 0;
for my $i (0 .. @terms-1) {
    $sum += $values[$i] if $terms[$i] ne '?';
}
print "$input = $sum\n";

    [download]

    #11929 First ask yourself `How would I do this without a computer?' Then have the computer do it the same way.

[reply]
[d/l]
[select]

      Using substr might be slightly faster than split-ing to an array (YMMV): 

      c:\@Work\Perl>perl -wMstrict -le
"my @vectors = (
   [ '--+-+', 122 ], [ '---++', 122 ], [ '?????', 0 ],
   [ '-????',   1 ], [ '-???-', 101 ],
   );
 ;;
 my @map = (1, 7, 11, 3, 100);
 ;;
 use Test::More 'no_plan';
 ;;
 for my $ar_vector (@vectors) {
   my ($s, $expected) = @$ar_vector;
   ;;
   my $cvt = convert($s);
   ok $cvt == $expected, qq{'$s' -> $expected (got $cvt)};
   }
 ;;
 sub convert {
   my ($s) = @_;
   ;;
   my $sum = 0;
   $sum += substr($s, $_, 1) ne '?' ? $map[$_] : 0 for 0 .. length($s)
+-1;
   return $sum;
   }
"
ok 1 - '--+-+' -> 122 (got 122)
ok 2 - '---++' -> 122 (got 122)
ok 3 - '?????' -> 0 (got 0)
ok 4 - '-????' -> 1 (got 1)
ok 5 - '-???-' -> 101 (got 101)
1..5

      [download]

[reply]
[d/l]
Re: decoding pattern expression
by Anonymous Monk on Aug 22, 2014 at 21:53 UTC

    This works for your sample data but probably not your real data:

    my $s = pack 'C5', qw[1 7 11 3 100];

say "$_: ", unpack '%C5', $s & tr/?+-/\0\xff/r for qw[
  --+-+
  ---++
  ?????
  -???+
];

    [download]

    Output:

    --+-+: 122
---++: 122
?????: 0
-???+: 101

    [download]
[reply]
[d/l]
[select]
Re: decoding pattern expression
by GotToBTru (Prior) on Aug 22, 2014 at 14:28 UTC

    What have you tried so far?

    1 Peter 4:10
[reply]
      i have tried a bit with bash scripting but didnt get too far
[reply]

        That tells me nothing. I suppose I should have asked before that, do you have to use bash scripting? You probably can do it that way but it really isn't intended for that. And, this is a perl forum, not a bash scripting forum, so you may have to look elsewhere. OTOH, perl is the perfect tool for this problem. Interested in learning something? Check a perl tutorial like the ones on this site, and post what you come up with. We'll help.

        1 Peter 4:10
[reply]
        FYI, that generally means:
        Show picturescode or it didn't happen.
[reply]

Back to Seekers of Perl Wisdom