Names to ids

David Landgren david at
Wed Jan 25 14:11:46 GMT 2006

Simon Wistow wrote:
> I need to turna filename into a numeric id in a repeatable manner (i.e 
> the same filename must always produce the same number).  

It doesn't matter if two file names map to the same number?

> My first thought (md5_hex the name and then hex the reuslt to get an 
> integer) gets an overflow error unless I truncate to the first 8 
> characters.

If you want all the bits to influence the outcome, then xor the 4 words 
of the MD5 checksum together. Something like

     my $number = 0;
     $number ^= $_ for unpack 'N4', md5($filename);

should do the trick.


> String::KoreMutake doesn't work because it can only taken certain 
> phonemes. 
> Any ideas?
> Simon

"It's overkill of course, but you can never have too much overkill."

More information about the mailing list