Comparing files line by line using a simple pattern match - regex

Question

I have two files: in the first file each line has some labels associated with it; the second file contains the labels which fall under certain categories.
File1 - labelled lines:
I have never had an issue. L_102 ----- L_127
I travel overseas and offer a lot of services that are very useful to me L_105 ----- L_134 ----- L_148
Expense to have L_522
Great benefits L_148
prestige L_118
File2 - categories under which the labels fall:
Issues:113,114,115,116,127
Benefits:105,220,154,543,590
General:148,134,154
I have written a Perl script to fetch labels from the first file.
#!/usr/bin/perl
use strict;
use warnings;
my $file = shift || "INPUTFILE";
my $outputfile = shift || "OUTPUTFILE";
open my $fh, '<', $file or die "Can not open '$file': $!";
open( OUTFILE, ">", $outputfile) or die "Can not open '$outputfile': $!";
while(my $w = <$fh>) {
my #matches = $w =~ m/(L_[0-9][0-9][0-9])/g;
for(#matches){s/L_//g;
s/\s+/\t/g;
}
print OUTFILE "#matches\n";
}
The output from this first script is:
102 127
105 134 148
522
148
118
I have a second Perl script to fetch the levels from second file (which contains the categories):
#!/usr/bin/perl
use strict;
use warnings;
my $file = shift || "INPUTFILE";
my $outputfile = shift || "OUTPUTFILE";
my $patern = shift ||"Issues:"
open my $fh, '<', $file or die "Can not open '$file': $!";
open( OUTFILE, ">", $outputfile) or die "Can not open '$outputfile': $!";
while(my $var = <$fh>) {
if(my #matches =$var=~/(.*$patern.*)/)
{
for(#matches){s/$patern//g;s/\,/\t/g}
print OUTFILE "#matches\n";
}
}
The second output from the second script is:
113 114 115 116 127
Now I want to match the first output with the second output line by line.
The results I want are: if the any of the numbers in the second output matches with any of the lines in the first output then I want to print 1; or else if there is no match print -1 for that line.
The output from the above would be as below:
1
-1
-1
-1
-1

Answer 1

This combines your two scripts into one. It reads through the $inputfile file handle that is pointing at "INPUTFILE.txt"looking for matches based on either a regular expression ($regexp) or the existence of a search key in the %patterns hash.
Since the match is simple, the regular expression we use is built up using join, |, and the required search strings. In the alternative approach (which is commented out here) we use the hash keys themselves to check whether a search pattern exists.
I have changed the variables and file names in the open statements somewhat since the capitalized file names made them seem like old style file handles:
#!perl -l
my $inputfile = "INPUTFILE.txt";
my $outputfile = "OUTPUTFILE.txt";
my $matchfile = "MATCHFILE.txt";
open my $inputfh, '<', $inputfile or die "No file '$inputfile': $!";
open my $matchfh, '<', $matchfile or die "No file '$matchfile': $!\n";
open my $outfh, '>', $outputfile or die "No file '$outputfile': $!\n";
my %patterns;
while (<$matchfh>) {
$patterns{$_} = () for map { split /,/, $_ } /Issues:(.*)/;
}
my $regex = join "|", keys %patterns;
$regex = qr/$regex/; # create a regex from %patterns
print "Search patterns : ", join " ", keys %patterns;
print "Regex : $regex \n";
while (my $line = <$inputfh>) {
chomp $line;
# Print "1" for 3 digits matching search pattern; "-1" otherwise:
#print exists $patterns{$_} ? "1" : "-1" for $line =~ m/(\d\d\d)/g;
# Print "1" if a matching pattern is on a line; -1 otherwise:
if (grep /$regex/, $line) { #
print "1 - $line";
}
else {
print "-1 - $line";
}
}
The above script should work. You can remove - $line from the last print statements and add a file handle destination ($outfh) to direct the output to a file.
Since there are five lines in the inputfile, the output is:
Search patterns : 127 116 114 115 113
Regex : (?^:127|116|114|115|113)
1 - I have never had an issue. L_102 ----- L_127
-1 - I travel overseas ... very useful to me L_105 ----- L_134 ----- L_148
-1 - Expense to have L_522
-1 - Great benefits L_148
-1 - prestige L_118
NB the final if ... else blocks could be shortened using the "ternary operator"(<cond> ? 1 : 0) to:
print $line =~ /$regex/ ? '1' : '-1';
so that "1" will printed if $line =~ /$regex/ evaluates to "true" (or "1") ; and "-1" will be printed if it evaluates to "false" (or "0").
If you read from your two files and simply redirect the output with your shell, the short version of all this would be:
#!perl -l
my $inputfile = "INPUTFILE.txt";
my $matchfile = "MATCHFILE.txt";
open my $inputfh, '<', $inputfile or die "No '$inputfile': $!";
open my $matchfh, '<', $matchfile or die "No '$matchfile': $!\n";
my %patterns;
while (<$matchfh>) {
$patterns{$_} = () for map { split /,/, $_ } /Issues:(.*)/;
}
my $regex = join "|", keys %patterns;
$regex = qr/$regex/;
while (my $line = <$inputfh>) {
chomp $line;
print $line =~ $regex ? '1' : '-1';
}