Shorten a regex, allowing false positives

regex_shortener.py

s = 'amsterdam|london|tokyo|indianapolis|new york|shanghai|toronto|san francisco'
rg = s.split('|')

# count: the length of the shortest word
count = len(min(rg, key=len))
 
# v: list of sets. v[0] is the set of all the first letters, etc.
v = []
for i in range(0,count):
  v.append(set(w[i] for w in rg))
 
# now generate a regex from v by combining the sets in v
def make_block(chars):
  return '[' + str.join('', chars) + ']'
print str.join('', map(make_block, v)) # '[ailnst][aehmon][adknsrw][ dionty][aegfony]'

Use case: You have a bunch of IDs in your Google Analytics that you want to match on. It's very long and doesn't fit anymore.

If false positives are OK, you can try to make it shorter by using character classes. Some false positives from this example: words starting with lenin, santa, toddy, nnnnn, thane, and 8562 other unintended prefixes. If that's acceptable, you can use this toy.