fastest way to test file for string?

[Scott David Daniels]

Tim Chase wrote:
>> Hi.  I need to implement, within a Python script, [functionality like]:
>>    grep -rl some_string some_directory
> 
> I'd do something like this untested function:
> 
>   def find_files_containing(base_dir, string_to_find):
>     for path, files, dirs in os.walk(base_dir):
Note order wrong here
>       ...
> 
>   for filename in find_files_containing(
>       "/path/to/wherever/",
>       "some_string"
>       ):
>     print filename

I like results in a nice order, so I do something more like:

     def find_files_containing(base_dir, string_to_find):
         for path, dirs, files in os.walk(base_dir): # note order
             for fname in sorted(files):  # often endswith in here
                 full_name = os.path.join(path, fname)
                 try:
                     with open(full_name) as f:
                         for line in f:
                             if string_to_find in line:
                                 yield full_name
                                 break
                 except IOError, why:
                     print ("On %s in %s: %s' % (fname, path, why))
             # usually several subdirs to avoid
             dirs[:] = sorted([d for d in dirs
                        if d[0] != '.' and d not in ('RCS', 'CVS')])

--Scott David Daniels
Scott.Daniels at Acm.Org