[Originally posted by mpwright]

Hello. I need to find and replace 2,000,000 numbers in 24 files each 150mb in
size. A normal s/// method works far to slowly. Do you know any modules or
tree search algorithms that could work on files of this size and not take a
week to run! Thanks, Matt.

[Originally posted by dave]

> Hello. I need to find and replace 2,000,000 numbers in 24 files each 150mb in
> size. A normal s/// method works far to slowly. Do you know any modules or
> tree search algorithms that could work on files of this size and not take a
> week to run! Thanks, Matt.

Matt,

The problem probably isn't with the s/// part of your code, but with the file
accesses. Are you reading the files a line at a time? That's going to be very
slow. Is there a larger record that you count read in at one time? Or could
you just read the file 500k at a time?

Take a look at the $/ variable (either in perldoc perlvar or on pp111-114 of
my book.

Let me know if you need any more help.

Dave...