I tried the two approaches, cant really spot a performance difference.
both take between 3 and 4 seconds on 1600 instances -
that is including a logmessage listing the instance and the master.
if tried a complete version in OM (the logic I described below), doing
several tests, with typename, amount of childs, groups, amount of owner,
typename of owners, untill all conditions were met for an instance. Ten or
so lines, about four conditions in there, two "for each" loops.
then I stripped everything down to the bare minimum, assuming that instances
were selected, immediately navigating to owners, finding a group in there
and going to the owner, one "for each" loop and one condition. I was
surprised to see that it didnt affect the speed noticeably.
And the command getmaster() was the same speed, however you would have to do
some error catching for when you run it on something which is no instance.
As far as determining if the model is an instance or a real model,
it'd be interesting to see how does your approach performs compred to
getmaster().
Cheers
Bernard
On Nov 20, 2007 4:27 AM, peter boeykens <peter_b(at)skynet.be> wrote:
> I never thought of checking the instance group. This returns the group
> under the source model, so I guess checking the parent of that group
> will return the source model of the instance. Nice one.
>
> Cheers
> Bernard
tried it out and got it to work, something like this:
test if its a model with .type
if it is, find childs and groups,
if both their counts are at 0,
find the owners
for each owner
test if its a group
if so, test the owner of the group
if its a pass: bingo! we have an
instance.
if it's a model: this is the master
object.
its a totally OM way, ran it on a hundred instances in almost no time.
---
Unsubscribe? Mail Majordomo(at)Softimage.COM with the following text in
body:
unsubscribe xsi
---
Unsubscribe? Mail Majordomo(at)Softimage.COM with the following text in body:
unsubscribe xsi
---
Unsubscribe? Mail Majordomo(at)Softimage.COM with the following text in body:
unsubscribe xsi
