On Fri, 2011-09-23 at 17:04 +0200, Nils Philippsen wrote:
> On Tue, 2011-09-20 at 11:29 -0400, James Antill wrote:
> > There's no need to make this a multiline list generator, just using a
> > for loop and append() is basically identical speed wise ... and stupid
> > people like me can more easily read it :).
>
> You'll notice that we really want to annoy people when we start using
> map() and filter(). :-P
My current feelings, which have changed over the last few years, is
that given:
1.
foo = []
for x in blah:
if y:
foo.append(x)
2.
foo = [x for x in blah if y]
3.
foo = filter(lambda x: y, blah)
...#1 is pretty much identical speed wise to the others, and everybody
can understand it easily. The only downside is that for the most simple
cases it has 2-4 times the vertically whitespace (but not much more
typing). So if there is _any_ doubt, just use #1.
#2 is fine when all of "x" "blah" and "y" are fairly simple. The limit
is probably something like:
return [pkg.pkgtup for pkg in self.returnPackages(patterns=patterns,
ignore_case=False)]
...if it's just the test that is "non-simple" you can always put it in a
function and use:
foo = [x for x in blah if y(...)]
...if it's just the "blah", you can always use a generator. Or, again, just use
a for loop.
#3 is fine when the first arg. is _naturally_ a function, and the
second is naturally a variable ... but just using a #2 is not a bad idea
anyway. We do not worship lisp :).
The general idea being that there is no reason to be "clever" for no
reason, esp. as we have to do that too often anyway.
_______________________________________________
Yum-devel mailing list
Yum-devel@lists.baseurl.org
http://lists.baseurl.org/mailman/listinfo/yum-devel
