Re: OT: hashing strings
- Posted by prickle Aug 05, 2009
- 1195 views
useless said...
What's your opinion of the fastest hash of 10million unique strings, each being 50char to 400char long, hashing down to a 32bit (or smaller) positive integer, preferably with unique hashes (or one result per bucket)? Preferably using no math.
The probability of a collision given any 32-bit key with a perfectly evenly distributed hash function on random data becomes 99.9972% at 300,000 unique strings. It's pretty much certain after that. Math Rocks! 
Cheers,
Nick
