[Python-Dev] SRE 0.9.8 benchmarks

[M.-A. Lemburg]

> 
> searching for literal text:
> 
> searching for "spam" in a string padded with "spaz" (1000 bytes on
> each side of the target):
> 
> string.find     0.112 ms
> sre8.search     0.059
> pre.search      0.122
> 
> unicode.find    0.130
> sre16.search    0.065
> 
> (yes, regular expressions can run faster than optimized C code -- as
> long as we don't take compilation time into account ;-)
> 
> same test, without any false matches:
> 
> string.find     0.035 ms
> sre8.search     0.050
> pre.search      0.116
> 
> unicode.find    0.031
> sre16.search    0.055

Those results are probably due to the fact that string.find
does a brute force search. If it would do a last match char
first search or even Boyer-Moore (this only pays off for long
search targets) then it should be a lot faster than [s|p]re.

Just for compares: would you mind running the search 
routines in mxTextTools on the same machine ?

import TextTools
TextTools.find(text, what)

-- 
Marc-Andre Lemburg
______________________________________________________________________
Business:                                      http://www.lemburg.com/
Python Pages:                           http://www.lemburg.com/python/