perl -i.bak -pe 's/frog/toad/g' txtfile
[download]

This one-liner will save your old file into txtfile.bak, and modify txtfile by changing every occurrence of "frog" into "toad".

See perlrun for details on the command line arguments.

As an addendum to dbwiz's solution: you may want to add word boundaries to the regexp if you want to avoid 'misfrogged' to be replaced by 'mistoadged' (ok, the example is contrived, but the potential problem isn't).

So the above would become:

perl -i.bak -pe 's/\bfrog\b/toad/g' txtfile
[download]

Hope this helps, -gjb-

gbj and dbwiz's solutions are more elegant, but if you want it inside an existing script, you could use something like this:

#!/usr/bin/perl

use warnings;
use strict;

my $file = "txtfile";

# Read the file into $old_text:
#
open my $old_fh, "<$file";
my $old_text;
do { local $/; $old_text = <$old_fh> };
close $old_fh;

# Do the substitution
#
( my $new_text = $old_text ) =~ s/\bfrog\b/toad/g;

# Write the new "toad" data out, overwriting the old file
#
open my $new_fh, ">$file";
print $new_fh $new_text;
close $new_fh;
[download]