in reply to Re: comparing columns and printing a result
in thread comparing columns and printing a result

This is an example of the source file: 
SUUTIN STAMM 2/60AST;SUUTIN STAMM 2,0/60 AST.           VIIRAOSA (PK-1
+)
SUUTINKÄRKI SPRAY TP 9501 E-SS;SUUTIN  VALME HOL0125624 TP 9501    E-S
+S
O-RENGAS 790X12 FPM;UPPOPUMPPU PUMPEX KV56
O-RENGAS 99X7 FPM;RULLAUSPÄÄ B NORMAALI
KÄSISAHA SANDV 2600-22-XT L=22IN;KÄSISAHA 22 IN 2600-22-XT          SA
+NDVIK
VEITSI STANL 10-010;MATTOVEITSI STANLEY 2-10-099 99E
VEITSENTERÄ STANL 11-916 L=62MM SUORA;MATTOVEITSENTERÄ  STANLEY 0-11-9
+21    (5KPL/PAK)  PITT. 49
KAAVAIN TIZIT 620-20 H10;TASOKAAVAIN SANDVIK 620-25

[download]
This is supposed to be the goal: 
SUUTIN STAMM 2/60AST;SUUTIN STAMM 2,0/60 AST.           VIIRAOSA (PK-1
+);3;2
SUUTINKÄRKI SPRAY TP 9501 E-SS;SUUTIN  VALME HOL0125624 TP 9501    E-S
+S;5;2
O-RENGAS 790X12 FPM;UPPOPUMPPU PUMPEX KV56;3;0
O-RENGAS 99X7 FPM;RULLAUSPÄÄ B NORMAALI;3;0
KÄSISAHA SANDV 2600-22-XT L=22IN;KÄSISAHA 22 IN 2600-22-XT          SA
+NDVIK;4;3
VEITSI STANL 10-010;MATTOVEITSI STANLEY 2-10-099 99E;3;2
VEITSENTERÄ STANL 11-916 L=62MM SUORA;MATTOVEITSENTERÄ  STANLEY 0-11-9
+21    (5KPL/PAK)  PITT. 49;5;2
KAAVAIN TIZIT 620-20 H10;TASOKAAVAIN SANDVIK 620-25;4;1

[download]
Structure is: "word1 word2 word3;bla hword1bla word2h blah;3(words in 1st column);2(words found a match from 2nd column)" Ok I made changes in the code which now looks like this: 
#!perl

use strict;
use warnings;

my $readfile = 'blah.csv';
my $writefile = 'bleh.csv';

open my $fh, "<", $readfile;

my $row = <$fh>;
my $found = 0;
my @cols;
my (@col1,@col2);
my @words = split(/\s+/,@col1);

open my $wfh, ">", $writefile or die "yikes: $!";

while (<$fh>) {
    tr /a-z/A-Z/;
    chomp;
    my @cols = split /\;/;
    push @col1, $cols[0];
    push @col2, $cols[1];
    my @words = @words + 1;

    if ( @col2 =~ m/$words[$_](\d+)/ ) {
        $found++;
    }

    print $wfh "$row';'@words';'$found";

    $found = 0;
    @words = 0;
}

[download]
Now I get report: 
Applying pattern match (m//) to @array will act on scalar(@array) at C
+:\blah\vertailu2.pl line 27.
Argument "LSAIDHA 2FA SFF ;ASD 2FA AASDA" isn't numeric in array eleme
+nt at C:\blah\vertailu2.pl line 27, <$fh>
Argument "3FASFL FAAL;DAOIADJAD" isn't numeric in array element at C:\
+blah\vertailu2.pl line 27, <$fh> line
Argument "ASFD ADD AD7A ALUYAD;ADLIHADBA A DADASFD DADD" isn't numeric
+ in array element at C:\blah\vertailu2.pl line 27, <$fh> line 4.

[download]
I'm guessing I should somehow define the program to handle both numbers and letters. Don't know how to do it though..

Replies are listed 'Best First'.
Re^3: comparing columns and printing a result
by johngg (Canon) on Jan 26, 2009 at 16:16 UTC

    I have taken a different approach to the hash based one of cdarke and have used regular expression matching instead. The regular expression is an alternation of the words found in $col1 and doing a global match against $col2 will find all matches. We are not interested in the text of the matches, just the number which is what the my $matches = () = ... construct achieves.

    Note that I'm only reading your data from a HEREDOC and writing to a variable just to keep everything inside the script on my system. Just substitute normal files if you use some of this code.

    use strict;
use warnings;

open my $inFH, q{<}, \ <<EOF or die qq{open: << HEREDOC: $!\n};
SUUTIN STAMM 2/60AST;SUUTIN STAMM 2,0/60 AST.           VIIRAOSA (PK-1
+)
SUUTINKÄRKI SPRAY TP 9501 E-SS;SUUTIN  VALME HOL0125624 TP 9501    E-S
+S
O-RENGAS 790X12 FPM;UPPOPUMPPU PUMPEX KV56
O-RENGAS 99X7 FPM;RULLAUSPÄÄ B NORMAALI
KÄSISAHA SANDV 2600-22-XT L=22IN;KÄSISAHA 22 IN 2600-22-XT          SA
+NDVIK
VEITSI STANL 10-010;MATTOVEITSI STANLEY 2-10-099 99E
VEITSENTERÄ STANL 11-916 L=62MM SUORA;MATTOVEITSENTERÄ  STANLEY 0-11-9
+21    (5KPL/PAK)  PITT. 49
KAAVAIN TIZIT 620-20 H10;TASOKAAVAIN SANDVIK 620-25
EOF

my $outFile;
open my $outFH, q{>}, \ $outFile
   or die qq{open: > \ $outFile: $!\n};

while( <$inFH> )
{
    chomp;
    my( $col1, $col2 ) = map uc, split m{;};
    my $rxCol1 = do
       {
           local $" = q{|};
           qr{@{ [ map quotemeta, split m{\s+}, $col1 ] }}
       };
    my @col1Words    = split m{\s+}, $col1;
    my $matches = () = $col2 =~ m{$rxCol1}g;
    print $outFH
       join( q{;}, $col1, $col2, scalar @col1Words, $matches ),
       qq{\n};
}

close $inFH or die qq{close: << HEREDOC: $!\n};
close $outFH or die qq{close: > \ $outFile: $!\n};

print $outFile;

    [download]

    The output.

    SUUTIN STAMM 2/60AST;SUUTIN STAMM 2,0/60 AST.           VIIRAOSA (PK-1
+);3;2
SUUTINKÄRKI SPRAY TP 9501 E-SS;SUUTIN  VALME HOL0125624 TP 9501    E-S
+S;5;3
O-RENGAS 790X12 FPM;UPPOPUMPPU PUMPEX KV56;3;0
O-RENGAS 99X7 FPM;RULLAUSPÄÄ B NORMAALI;3;0
KÄSISAHA SANDV 2600-22-XT L=22IN;KÄSISAHA 22 IN 2600-22-XT          SA
+NDVIK;4;3
VEITSI STANL 10-010;MATTOVEITSI STANLEY 2-10-099 99E;3;2
VEITSENTERÄ STANL 11-916 L=62MM SUORA;MATTOVEITSENTERÄ  STANLEY 0-11-9
+21    (5KPL/PAK)  PITT. 49;5;2
KAAVAIN TIZIT 620-20 H10;TASOKAAVAIN SANDVIK 620-25;4;1

    [download]

    I hope this is helpful.

    Cheers,

    JohnGG

    Update: I just noticed the bit about pipes in your data so added a quotemeta in the regex.

[reply]
[d/l]
[select]

In Section Seekers of Perl Wisdom