http://www.perlmonks.org?node_id=1052905

Mark.Allan has asked for the wisdom of the Perl Monks concerning the following question:

perl version = v5.8.2

I have an issue with parsing a file, I have for instance 5 keys in a %results hash which I have already populated key1,key2,key3,key4,key5

Using them key values, I need to open a file, search for the line which matches each key, then process the attribute values which exist before the ENDKEY for each key and update the %results hash. for example:

HIT key1 so start to process attrib1: base = FULL attribmsg: middle = MEDIUM ENDKEY HIT key2 so start to process attrib1: base = FULL ENDKEY HIT key9 so start to process attrib1: base = FULL ENDKEY

Notice key9 is in the file but is not in my hash therefore its keys and data attributes should never be matched.

snipet of code so far which dont work

open(FH,"<$file") || die ("cannot open file"); foreach my $key (sort keys %results){ local $/ = "ENDKEY\n"; while (my $line =~ /$key/){ if ($line =~ /\S+\s($key).*/){ my $keymatch = $1; $results{$keymatch}{attrib1} = 'null'; $results{$keymatch}{attribmsg} = 'null'; } if (/\s+(attrib1|attribmsg):.*=\s(\S+)/){ $results{$keymatch}{$1} = $2;} } }

Replies are listed 'Best First'.
Re: Matching hash keys in text file
by Eily (Monsignor) on Sep 08, 2013 at 13:46 UTC

    What's wrong with your code:

    • You never read your input file, you just open it. There should either be a readline somewhere or a <FH>
    • The =~ operator binds the variable on its left to the expression on its right. So in my $line =~ /$key/ you try to find $key in a brand new variable with nothing in it (that's what my does). Obviously you don't find it. Just below that, you have a lonely (/\S+\s($key).*/; there's no =~ so you do that on the default variable : $_ . Since you never wrote in it, that's probably not what you meant.
    • Your first $1 contains the same thing as $key, since that's what you were looking for in the first place.
    • It's probably better to work the other way around: loop through the file once and for each paragraph search through all the keys. Or, search for a key in paragraph and check that it exists
    • There's a way to read a file paragraph by paragraph, you just have to set $/ to "". But what you did you still work

    From all of that:

    use strict; use warnings; use Data::Dumper; my %result = (key1 => undef, key2 => undef, key3 => undef); { # So that the effect of local is limited to that block local $/ = 'ENDKEY'; while( my $line = <DATA>) # <FH> in your case { next unless $line =~ m<HIT \s+ (key\d+)\b >x; my $key = $1; next unless exists $result{$key}; $result{$key} = { attrib1 => 'null', attribmsg => 'nul +l' }; while ($line =~ m<\s+ (attrib1 | attribmsg) :.*? = \s+ + (\w+) >gx) { $result{$key}{$1} = $2; } } } print Dumper \%result; __DATA__ HIT key1 so start to process attrib1: base = FULL attribmsg: middle = MEDIUM ENDKEY HIT key2 so start to process attrib1: base = FULL ENDKEY HIT key9 so start to process attrib1: base = FULL ENDKEY
    $VAR1 = { 'key2' => { 'attribmsg' => 'null', 'attrib1' => 'FULL' }, 'key1' => { 'attribmsg' => 'MEDIUM', 'attrib1' => 'FULL' }, 'key3' => undef };