What part of the replies you got to Comparing a Hash key with a variable (if statement) have you applied?

Also, if this is all you have problems with:

My problem: If the values in col 'a' are <= .03, I need to print the entire row of the input file to the output file, not just the value from the column I am working with. Any help with this issue would be much appreciated. Thanks.

... why do you post the large program which has very little to do with your problem? Reduce your program, for example by hardcoding the filters and filter conditions.

Also, you might want to look at your print statement and just print out all you want instead of only printing the matched cell.

While I do not know the answer to your current question, I do think your code is irreparably broken as long as you insist on using eval like that. BrowserUK gave a very good suggestion here to invoke the operators using a dispatch table (a hash of function references, each taking two arguments). You should incorporate that into your code.

Anyway, eval is so very rarely needed and should be shunned. But you wanted minor corrections. Let's see...

Change

if ((eval "$cells[$extract_col] $filter_array[$i][1] $filter_array[$i]
+[2]"))
[download]

to

if ((eval '$cells[$extract_col] ' . $filter_array[$i][1] . ' $filter_a
+rray[$i][2]'))
[download]

I need to print the entire row of the input file to the output file

Then why don't you do so? Change

print OUTFILE "$cells[$extract_col]", "\n";
[download]

to

print OUTFILE $row, "\n";
[download]

The clarity of your code would be somewhat enhanced if you gave @filter_array's columns names; e.g. $filter_array[$i]->{action} = 'append' or $filter_array[$i]->{column} = 'c'. Remember: with clear code, it is easier to be certain that it works correct.

Oh, and please stop quoting "$var" since it is identical to plain $var -- barring a few exceptions having to do with references and objects.

After extracting your filter preferences, you start a first ("outer") loop where you go through the filters one by one.

Inside that loop you read the input file and check the filter condition for each line of the input file ("inner" loop).

BUT when you end the "inner" loop you have exhausted the input file and when you start the next iteration of your "outer" loop, <IN> will return nothing and your "inner" loop gets skipped.

I really doubt that is what you meant to do. Rather you should open the input file inside the outer loop and close it again after the "inner" loop is finished. The next iteration of the "outer" loop will then open the input file again and start reading it afresh and checking your next filter. Of course that is a slow and wasteful way of working: for every filter you walk the input file afresh. It is OK when the files are small, but it quickly will become unwieldly and slow when the files grow.

Look at the following:

use Modern::Perl;
use Data::Dump qw/dump/;
use SQL::Abstract;
use DBI;

my $dbh = DBI->connect(
    'dbi:CSV:',
    '', '',
    {
        f_dir        => './DB/',
        csv_sep_char => "\t",
    }
) or die "Cannot connect: $DBI::errstr";

my @where;
{
    my $query = 'SELECT * FROM filter_file ORDER BY orde';
    my $sth   = $dbh->prepare($query);
    $sth->execute();

    while ( my $row = $sth->fetchrow_hashref ) {
        push @where,
          { $row->{'column'} => { $row->{'relationship'} => $row->{'va
+lue'} } };
    }
    $sth->finish();
}
my $sql = SQL::Abstract->new;
my ( $where, @bind ) = $sql->where( \@where );

{
    my $query = "SELECT a, b, c FROM input $where";
    my $sth   = $dbh->prepare($query);
    $sth->execute(@bind);

    while ( my $row = $sth->fetchrow_hashref ) {
        say dump($row);
    }
    $sth->finish();
}
[download]

{ a => 0.2, b => "abc", c => 0.3 }
{ a => 0.1, b => "abd", c => 0.3 }
{ a => 0.4, b => "abe", c => 0.2 }
{ a => 0.1, b => "abc", c => 0.5 }
{ a => 0.1, b => "abd", c => 0.8 }
[download]

No repeated going through your input file; clean and fast code; easy to understand, maintain and change for anyone who knows the basics of SQL.

A few caveats though:

• I had to change the name of your order column to orde since "order" is a reserved SQL keyword.
• As we are now dealing with SQL, the "equals" test is always =, even for strings so I had to change your eq in the filter_file to =.