speed up perl hashes in foreach loop and better algorithm - perl

Question

I have 2 hashes -> %a and %b.
Hash %a is from temp.txt
my %a = map{
my $short = substr($_,12);
$count++ => {$short => $_};
} #a;
my %b = map {
$_ => $_;
} #b;
%a = (
'1' => {'We go lunch' => 'We go lunch 9 pm'},
'2' => {'We go break' => 'We go break 8 pm'},
'3' => {'We go lunchy' => 'We go lunchy 8 pm'}
);
%b = (
'We go lunch' => 'We go lunch',
'We go break' => 'We go break',
'We go lunchy' => 'We go lunchy'
);
foreach my $key (keys %a){
foreach my $key2 (keys %{$a{$key}}){
if(exists $b{$key2}){
delete $a{$key}{$key2};
delete $a{$key};
}
}
}
my #another;
foreach my $key ( sort {$a<=>$b} keys %a) {
foreach my $key2 (keys %{$a{$key}}){
$another[$count] = $a{$key}{$key2};
$count++;
}
}
how can I speed up this? is my hashes weird? It took 30 seconds to output #another through 25144 lines of words in temp.txt.
Is it necessary to make hash of hash for %a?
The reason is I want any %b{$key} value in %a to be deleted.
I'm still learning Perl if you guys have any better way of doing this, very much appreciated, possibly using map and grep? and better algorithm?
previous workaround
if you see every #b is shorter string than every #a but still within #a. I had try to use
foreach (#b) {
my $source = $_;
#another = grep !(/$source/i), #a;}
but still it doesn't work. I was confused and thus came this hash of hash in %a and make hash %b from #b just to get rid of every instances value of #b in #a. which comes out as weird hash. lol

Answer 1

There are a few unknowns here - how is %b built for example.
Otherwise, a few observations:
You should use another array instead of %a:
my #c = map{
{ "".substr($_,12) => $_}
} #a;
If you already have %b defined, you could further optimize it by:
my #another = grep !exists $b{ substr($_,12) }, #a;
Hope this helps
Also, don't forget to always use strict; and use warnings; in the beginning of your program.
Explanations:
Your code puts everything in %a, traverse it and eliminates what shouldn't be there.
I think you could simply grep and keep in an array only the desired results.
The optimized code should become:
use strict;
use warning;
my %b = (
'We go lunch' => 'We go lunch',
'We go break' => 'We go break',
'We go lunch' => 'We go lunch'
);
#add code that initially fills #a
my #another = grep { !exists $b{ substr($_,12) } } #a;

Answer 2

It seems you are very confused. First of all, substr $_, 12 returns all the characters after the twelfth in the string, and so doesn't create the data structure you say it does. Secondly, you are using hash of hashes %a as an array of arrays, as the keys are integers without gaps in the sequence, and the value you are storing is a simple string pair.
The biggest problem for us is that you don't explain your goal in all this.
What it looks like is that you want to end up with the array #another containing all the lines from temp.txt that don't begin with any of the strings in #b. Is that about right?
I would do it by building a regular expression from array #b, and checking each line from the file as I read it.
This program demonstrates. I have renamed array #b to #exclude as the former is a terrible name for a variable. The regular expression is built by preceding each element of the array with ^ to anchor the regex at the beginning of the string, and appending \b to force a word boundary (so that, for instance, lunch doesn't match lunchy). Then all the elements are joined together using the | alternation operator, resulting in a regex that matches a string that starts with any of the lines in #exclude.
After that it is a simple matter to read through the file, check each line against the regex, and push onto #another those lines that don't match.
Note that, as it stands, the program reads from the DATA file handle so that I could include some test data in the source. You should change it by uncommenting the open line, and deleting the line my $fh = *DATA.
use strict;
use warnings;
#open my $fh, '<', 'temp.txt' or die $!;
my $fh = *DATA;
my #exclude = (
'We go lunch',
'We go lunchy',
'We go break',
);
my $exclude_re = join '|', map "^$_\\b", #exclude;
my #another;
while (my $line = <$fh>) {
chomp $line;
push #another, $line unless $line =~ $exclude_re;
}
print "$_\n" for #another;
__DATA__
We go breakfast 6 am
We go lunch 9 pm
We go break 8 pm
We go lunchy 8 pm
We go supper 7 pm
output
We go breakfast 6 am
We go supper 7 pm