Comments on: An Anagram Detector

By: Shriphani’s Weblog - Huge Anagram List

Shriphani’s Weblog - Huge Anagram List — Thu, 21 Aug 2008 22:51:08 +0000

[...] got on this blog) over to my account. Then, I ran a very huge file full of words through my anagram detector and the results were astonishing, a record 1.93 seconds to parse over 10000 words and pick [...]

By: Maarten van Emden

Maarten van Emden — Sat, 12 Jul 2008 19:00:02 +0000

I found your blog when searching for the source of the anagram program I vaguely remembered. So it was Bentley. Thanks. I also found http://pramode.net/2005/11/18/teaching-tips-the-anagram-problem-solved-in-unix-style/
Programming it is fun, but it is also fun to do it without programming, which is what Bentley showed. Bentley tells how Tom Cargill (another person worth knowing) describes the method: “first sort this way (moves hand from left to right), then sort that way (moves hand up and down).

By: Justin

Justin — Tue, 24 Jun 2008 19:15:06 +0000

def sign(word):
#can’t hash a list
return tuple(sorted(word))

def anagramCollect(filename):
f = open(filename, “r”)
ana_dict = {}
for word in f:
word = word.rstrip()
signature = sign(word)
ana_dict.setdefault(signature,[]).append(word)
f.close()
return ana_dict

By: J

Sun, 22 Jun 2008 02:38:41 +0000

To be even lazy, we can do:

from itertools import groupby

def sign(xs):
return “”.join(x + str(len(list(g)))
for x, g in groupby(sorted(xs)))

By: Ben

Ben — Mon, 16 Jun 2008 11:47:16 +0000

@DiThi: good call.

If you’re using a recent Python, consider just doing: “”.join(sorted(WORD)).

By: DiThi

DiThi — Mon, 16 Jun 2008 02:58:46 +0000

Here’s my signature maker:

def sign(word):
x=list(word)
x.sort()
return ”.join(x)

It’s a different format of signature but I’m sure it’s faster.

By: Mike Pirnat

Mike Pirnat — Sat, 14 Jun 2008 12:46:41 +0000

To get around the repeated slicing of the last character in your word, you could always do:

for word in f:
word = word.strip()
…