Creating CSV term document matrix from a hash stored in multideminsional array

[lobs]

PROBLEM AT HAND: every row in my csv file has the same value. I am trying to implement a term document matrix and store it in a CSV file.In my program I store the features and the value within a hash for a document. Each hash is stored in an array. The array then is stored within another array to seperate the document classes. I get the hash function by going through an array and adding the terms to the hash as follows (hash %termFreq is specific for the document and hash %docTerms is the features of all documents):

           
 while ($element = shift(@numOfWordArr)) {
              $termsFreq{$element} ++;
              if(!exists($docTerms{$element}))
              {
                $docTerms{$element}++;
              }
}

Then add the hash into an array of documents as follows:

push(@docArray, \%termsFreq);

Finally pass the array into an array to seperate the classes of the documents:

push(@classArr, \@docArray);

took the advice from another thread to iterate through the arrays to get a hash value and print the doc matrix to a csv file:

foreach my $subArr_ref(@classArr) {
  print "subArr_ref: ".@{$subArr_ref}."\n";
  foreach my $hashRef(@{$subArr_ref}) { 
    print "hashRef: ".$hashRef."\n";
    foreach my $key (sort keys %{$hashRef}) {
      #print $csv $key.":".$hashRef->{$key}.",";
      print $csv "$key : ${$hashRef}{$key},";
    }
    # foreach my $feat(@featureVector) {
    #   print $csv $hashRef->{$feat}.",";
    # }
    print $csv $i."\n";
  }

My problem is that I get rows of the same value since it is accessing the same hash per iteration. Help is much appreciated. Thanks!

[stevieb]

This:

for my $i($#classArr) {

Does not increment $i like you think it does. Essentially, it loops only once, because $#classArr is only a single number; the last element number in the array.

You want something more like:

for my $i (0..$#classArr){

Which will put the current element number into $i on each iteration, starting from 0.

update: Or, more idiomatically (untested):

for my $class (@classArr){
    for my $href (@$class){
        for my $key (sort keys %$href){
            print $csv "$key-- $href->{$key},";
        }
    }
}

[lobs]

yea I changed it recently and its now:

foreach my $subArr_ref(@classArr) {
  print "subArr_ref: ".@{$subArr_ref}."\n";
  foreach my $hashRef(@{$subArr_ref}) { 
    print "hashRef: ".$hashRef."\n";
    foreach my $key (sort keys %{$hashRef}) {
      #print $csv $key.":".$hashRef->{$key}.",";
      print $csv "$key : ${$hashRef}{$key},";
    }
    # foreach my $feat(@featureVector) {
    #   print $csv $hashRef->{$feat}.",";
    # }
    print $csv $i."\n";
  }
}

[stevieb]

Did that work?

Note I updated my original reply before you posted this update, but it would be more idiomatic Perl to rewrite it something like this (untested):

for my $class (@classArr){
    for my $href (@$class){
        for my $key (sort keys %$href){
            print $csv "$key-- $href->{$key},";
        }
    }
}

[Cow1337killr]

I discovered Text::CSV, recently (because I hang out at PerlMonks). Also, DBD::CSV let's one use SQL on their CSV files.