I'm solving some problem that involves Rabin–Karp string search algorithm. This algorithm requires rolling hash to be faster then naive search. This article describes how to implement rolling hash. I implemented "Rabin-Karp rolling hash" without problems and found few implementations implementations, but article also mentions computational complexity and that hashing n-grams by cyclic polynomials is prefered. It links to BuzHash imp开发者_如何学Clementation of such technique but I wonder how it can be used to build n-gram hash on top of it. I want to have something like this or
CPHash cp = new CPHash("efghijk");
cp.shiftRight('l') // now we got hash of "fghijki"
cp.shiftLeft('d') // "defghi"
for java.
For people who will encounter problems related with string search (like me) there are some articles that I found usefull: 1, 2, 3
I recently published an Apache licensed Java library which implements several rolling hash functions including Cyclic and Rabin-Karp:
http://code.google.com/p/rollinghashjava/
https://github.com/lemire/rollinghashjava
精彩评论