http://www.perlmonks.org?node_id=652132


in reply to Text difference

The problem is the size of the file is large(About 800Mb to 2Gb)
Repeating what was discussed in the CB, tye maintains Algorithm::Diff so he'd be your best bet for issues related to the module.

With such large files though, you may need to change your approach. Assuming that Algorithm::Diff is bogging down due to the large file size, the idea is that you should make sure that not so much of the data is loaded at one time.

Method 1

Algorithm::Diff works from two arrays, right? If so perhaps the easiest thing to do is just to pass it (references to) two Tie::File objects.

Update: well, it's not that simple. Algorithm::Diff builds a hash to keep track of indexes, which grows (O(?)) with the length of the arrays passed.

Method 2

If that didn't work (or was too slow) I'd do something like this (pseudocode):

declare @differences open file1 and file2 read chunk1 from file1 (64mbyte) read chunk2 from file2 (64mbyte) while (chunk1 contains data) compute h1: md5hex of chunk1 compute h2: md5hex of chunk2 if h1 != h2 compute @diff: pass chunk1 and chunk2 to Algorithm::Diff append @diff to @differences end-of-if read chunk1 from file1 (64mbyte) read chunk2 from file2 (64mbyte) end-of-while close file1 and file2 #maybe combine adjacent deltas in @differences (eg: by line number) report @differences
That way Algorithm::Diff only ever deals with 2x 64Mbyte of data at a time. If two adjacent chunks both contain differences between the files, you might investigate whether there's some way to combine those differences.

-David