I would go about it a little different. I would read your find/replace table into a hash with the find value as the key of the hash. Depending on the amount of memory on your box you are running this with, you may need to tie this has to a file in order to accommodate the data. With find/replace type stuff, I have found better performance with DB_BTREE instead of DB_HASH. Using this approach on a 4 processor 4GB ram server, I have processed close to 12 million an hour. By doing this way you would have something like:

#pseudo code
open lookup file
while(<lookup>) {
chomp;
my ($find,$replace)=split /=/;
$hash{$find}=$replace;
}
close lookup

open file to work on
while(file) {

my $field_to_lookup_on=split/unpack whatever

if(exists $hash{$field_to_lookup}) {
$new_field=$hash{$field_to_lookup};
}
else {
$new_field=$field_to_lookup_on;
}

put field back in record
print record.
[download]

This method does away with looping through arrays and puts it into a little better of a data structure. let me know if this doesn't make sense. kudos. krazken