While I don't have experience with this situation directly, I don't know
that it would end up making a whole lot of difference.
Both are C-types, though struct_time is 3 times the size of a float in
memory if that carries any weight in your situation. Either way, dicts are
efficient enough that I really don't think you'll run into a problem,
especially since I don't think the size of value objects affects lookup
time.
However, I don't want to assume too much about your use-case. Can you
provide any info on what you'll end up doing with these time values, how
often you update/query the dict and under what circumstances, and whether
you're worried about speed or memory efficiency (or both)? For example, if
you're going to be querying them a lot and then having to re-cast them to a
datetime or struct_time object, you'd probably be better off storing them as
such to begin with. Also, to what (rough) upper bound would you like to
bulletproof this? 250,000? 1 million? 25?
-Nathan
-----Original Message-----
From: Frank Rueter
Sent: Monday, June 11, 2012 11:37 PM
To: Nuke Python discussion
Subject: [Nuke-python] performance question
hm, that subject line might get stuck in some people's spam filter. oh
well...
I am storing hundreds of struct_time objects in a fairly large
dictionary and am wondering if it would be wiser to convert those to a
floating point numbers via time.mktime() first for efficiency?! The
dictionary is representing database content so it may grow a lot larger
in the future, which is why I'm keen to be efficient now rather than fix
things later.
Does anybody have experience wit this?
Cheers,
frank
_______________________________________________
Nuke-python mailing list
Nuke-python@support.thefoundry.co.uk, http://forums.thefoundry.co.uk/
http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-python
_______________________________________________
Nuke-python mailing list
Nuke-python@support.thefoundry.co.uk, http://forums.thefoundry.co.uk/
http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-python
