List comprehensions and glob

[Duncan Booth]

Peter Otten <__peter__ at web.de> wrote in news:cd5gvg$l4u$06$1 at news.t-
online.com:

> Duncan Booth wrote:
> 
>> If you want a list comprehension then how about:
>> 
>> filenames = [ name for pat in PATTERNS for name in glob.glob(pat) ]
> 
>>>> PATTERNS = ["*t", "*xt"]
>>>> filenames = [name for pat in PATTERNS for name in glob.glob(pat)]
>>>> len(filenames), len(sets.Set(filenames))
> (66, 39)
> 
> Names that match more than one pattern will appear more than once in the
> list.
> 
As indeed they do in the original posting. I assumed that either Robin 
wanted that behaviour or knew it wasn't relevant (the example given had 
non-overlapping patterns).

To remove duplicates just pass through a dictionary or a set:

filenames = dict.fromkeys(filenames).keys()