Trexgreen
asked on
Creating a Matrix in perl...
I'm doing a computation and that will result in 3 variable (Document name, document value, document term frequency) The values of the computation is in 3 separate variable, so I was thinking of grouping the values into an array or hash so I can sort the result of the computation by the following rules.
The highest document value and the highest term frequency.
here is a sample data
doc1, 0.002, 1
doc2, 0.012, 2
doc3, 0.032, 1
the search Martrix should be like this
doc2, 0.012, 2
doc3, 0.032, 1
doc1, 0.002, 1
How can I create a matrix in perl? or an array of array?
The highest document value and the highest term frequency.
here is a sample data
doc1, 0.002, 1
doc2, 0.012, 2
doc3, 0.032, 1
the search Martrix should be like this
doc2, 0.012, 2
doc3, 0.032, 1
doc1, 0.002, 1
How can I create a matrix in perl? or an array of array?
ASKER CERTIFIED SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
ASKER
can I sort two columns at once?
I want to have the highes freq and the value
I want to have the highes freq and the value
ASKER
Just found this... I guess I can use this to sort the result
http://search.cpan.org/~dcantrell/Sort-MultipleFields-1.0/lib/Sort/MultipleFields.pm
http://search.cpan.org/~dcantrell/Sort-MultipleFields-1.0/lib/Sort/MultipleFields.pm
You can, it depends really on what you want. The <=> and cmp operators compare 2 values and return -1, 0 or 1. But you can write a sort function that makes it clearer.
So basically comparing the 1st column, if they are not equivalent, return the sort order just based on 1st column, but if they are equivalent, then return sort based on 2nd.
There may be an even more elegant solution to this, but this is how I would do it.
Keep in mind this is sorting on the fly. If you want to maintain a sorted datastructure with many items, you probably should sort them as you store them.
So basically comparing the 1st column, if they are not equivalent, return the sort order just based on 1st column, but if they are equivalent, then return sort based on 2nd.
There may be an even more elegant solution to this, but this is how I would do it.
Keep in mind this is sorting on the fly. If you want to maintain a sorted datastructure with many items, you probably should sort them as you store them.
#!/usr/bin/perl
my %docs;
$docs{doc1} = { value => 0.002, freq => 1 };
$docs{doc2} = { value => 0.012, freq => 2 };
$docs{doc3} = { value => 0.032, freq => 1 };
$docs{doc4} = { value => 0.032, freq => 2 };
print "Unsorted\n";
foreach my $doc (keys %docs) {
print "$doc $docs{$doc}->{value} $docs{$doc}->{freq}\n";
}
print "Sorted\n";
foreach my $doc (sort sort2desc keys %docs) {
print "$doc $docs{$doc}->{value} $docs{$doc}->{freq}\n";
}
sub sort2desc {
if($docs{$b}->{freq} > $docs{$a}->{freq}) {
return 1;
}
elsif($docs{$b}->{freq} < $docs{$a}->{freq}) {
return -1;
}
else {
# 1st field equal, so sort by 2nd field
return $docs{$b}->{value} <=> $docs{$a}->{value};
}
}
$docs{$b}->{freq} <=> $docs{$a}->{freq}
to use whichever column to sort by. Note $b is before $a because you wanted descending sort.