djb / dbj2 hash in awk

djb2hash.awk

function ord(c) {
  for (i = 0; i < 256; i += 1) {
    if (sprintf("%c", i) == c) {
      return i
    }
  }
  return 256
}

function hash(str) {
  h = 5381
  for (pos = 1; pos <= length(str); pos += 1) {
    c = ord(substr(str, pos, 1))
    h = (((h * 33) + c) % 4294967296)
  }
  return h
}

{
  print hash($0)
}

Does the world need this? Probably not.

$ awk -f hash.awk <<EOD
> h
> he
> hel
> hell
> hello
> EOD
177677
5863442
193493694
2090324714
261238937

PS: This is a very slow implementation.

it's only slow cuz the full list of bytes are sequentially printed out in order to check. have it pre-made into array, or string lookup, and bunch up multiple characters before 1 mod and store op, and that should help a bit