Quick/East perl question

There are two sql statements/storedprocs. When executed, they return those two resultsets

sql1:  column1 column2 column3 column4
sql2:  column5 column3  
note: column3 is the same column for sql1 and sql2

 
in perl,

I have the code to get resultset from db as:

for sql1 query:

...
my (@lines, @data);
...
while( $databasehandle->dbresults != NO_MORE_RESULTS)
{
         $delimiter = "\t";
          for (;;)
         {
             ..........
             @data = $databasehandle->dbnextrow;
             push(@lines, join("$delimiter", @data));
            .........
         }
}
....
foreach $str (@lines)                 
{
        ($perlcol1, $perlcol2, $perlcol3, $perlcol4) = split /\t/, $str;
}

So this maps column1->$perlcol1 ...etc  column4->$perlcol4

Now i have the exact same code to run sql2 query except

foreach $str (@arr)
{

            ($perlcol5, $perlcol1) = split /\t/, $str;
}

this maps column5->$perlcol5,  and column3 -> $perlcol1

Note the difference between sql1 and sql2's result mapping
column1->$perlcol1
column3->$perlcol1

This code has been in production for a while and works fine so i assume it is correct, however it just doesnt make sense and seems like a bug. The program then uses $perlcol1 and $perlcol3 in a lot of hash tables etc.

For example, there is a hashtable defined as hashname{column1}, then it does
hashname{$perlcol1} to use it. EXCEPT now $perlcol1 points to column3 not column1. I dont see how they could possiblly map to the correct values. Am i right, or is it some perl syntax i am missing.

Thank you,

gaga