Question

Today I tried to cache re.compile to speed up my testing. Something got weird:

I used three cached re.compile:

re_cache1=re.compile(regexp_string1)
re_cache2=re.compile(regexp_string2)
re_cache3=re.compile(regexp_string3)
def mytest():
  m = re_cache1.split(data)
  if m:
     for item in m:
       m2 = re_cache2.search(data2)
       if m2:
         m3 = re_cache3.search(data3)

The result is not expected, m2 is None. But if I change a little bit, I will get m3’s match.

       m2 = re.search(regexp_string2, data2)

Anybody helps me out? Thanks very much~

Answer 1

Keep in mind: the re module caches internally anyway. The difference between:

re_cache1 = re.compile(regexp_string1)
...
re_cache1.search(data)

and

re.search(regexp_string1, data)

is one dictionary lookup. It’s usually not worth the extra bookkeeping to cache your own compiled regexes.