DJB2 Hash in Python

hash_djb2.py

#!/usr/bin/env python
# encoding: utf-8

def hash_djb2(s):                                                                                                                                
    hash = 5381
    for x in s:
        hash = (( hash << 5) + hash) + ord(x)
    return hash & 0xFFFFFFFF


hex(hash_djb2(u'hello world, 世界'))  # '0xa6bd702fL'

This is very slow and inefficient for long strings. You need to do & 0xFFFFFFFF on every step.

why & on every lap?

speaking of which, why there is & 0xFF'FF'FF'FF anyway? to make it uint32_t ?

@nmmmnu , yes. Because Python has bignum arithmetic by default. If you don't truncate the integer to uint32 on every step, the hash value will grow indefinitely and take longer and longer CPU time to process.

On top of that, DJB2 is not a good hash. FNV-1a is way better while being simple and similar in performance. See: https://gist.github.com/amakukha/7854a3e910cb5866b53bf4b2af1af968