Forums Technical Discussion Expression in Group SOP, strange results

Expression in Group SOP, strange results

   5552   9   1
User Avatar
Fr3drik
Member
36 posts
Joined: 10月 2007
Offline
 
Hi,

I have some trouble with the Group SOP behaving strangely.

I'm trying to group some primitives using the expression $TY >= $CEY && $TZ == $ZMAX on points, and then convert type to primitives.

It works as expected most of the time, but when I change the number of rows/cols in the initial grid to certain values, odd things happen - not all prims that should be are selected. Examples of values that generate this are (rows/cols) 2/8, 3/8, 4/6 (+ many more).

Would really appreciate some help, I just can't figure it out…

Thanks,
Fredrik

Ps. Using 10.0.374

Attachments:
group_expression.hipnc (39.2 KB)
4,6.png (25.5 KB)
5,6.png (26.0 KB)

 Report
User Avatar
aiworks
Member
130 posts
Joined: 4月 2009
Offline
 
Hmmm…. I wonder… seems to be working fine for me - see attached, is this the setup you used?

Cheers,
Greg

Attachments:
grouping-points.png (35.9 KB)

 Report
User Avatar
Fr3drik
Member
36 posts
Joined: 10月 2007
Offline
 
No, I just want to group the top row (as in my posted ‘img 5,6.png’).

That's why I group the points first and then convert the selection to prims (with 'Only include primitives with all points in the group). Please let me know if there is a better way of achieving this.

You see what I mean if you open the scene and try some of the row/col combinations I mentioned below; suddenly not all prims are selected.

Thanks,
Fredrik
 Report
User Avatar
aiworks
Member
130 posts
Joined: 4月 2009
Offline
 
I wonder if it might be a rounding bug in Houdini? Sometimes it miscalculates the TY for some points (see image), and works fine if you do the following: ($TY >= $CEY-.0001) && ($TZ >= $ZMAX)

Greg

Attachments:
grouping-points2.png (13.5 KB)

 Report
User Avatar
Fr3drik
Member
36 posts
Joined: 10月 2007
Offline
 
Thank you, does the job.

Yes, it does seem rather bug-ish to me. Being a novice, I didn't want to start yelling “bug”, but if you/someone could confirm perhaps it should be posted as one.

Is there usually a need for tolerance in selection situations like this?

Thanks,
Fredrik
 Report
User Avatar
aiworks
Member
130 posts
Joined: 4月 2009
Offline
 
Hi Fredrik,

Glad it works for you :-)

Beeing too newbie, too, I'll leave the “bug or not” issue to more experienced ones ;-) - is there any “Adept” or “Master” around to say more about it?

Greg
 Report
User Avatar
Fr3drik
Member
36 posts
Joined: 10月 2007
Offline
 
Ran in to a new problem: similar thing happens when I want to group the “center points” with $TY == $CEY (see img, not all points included with certain row/col combinations in Grid SOP).

I fixed it by using $TY >= $CEY-.0001 && $TY <= $CEY+.0001, but it is still annoying.

Is this a bug or am I missing something really obvious?

/Fredrik

Attachments:
group_points.jpg (131.9 KB)

 Report
User Avatar
aiworks
Member
130 posts
Joined: 4月 2009
Offline
 
Hi,

I'm surprised that you are surprised - if there is a rounding issue with $CEY, it will show wherever you try to use it ;-)

Cheers,
Greg
 Report
User Avatar
Fr3drik
Member
36 posts
Joined: 10月 2007
Offline
 
Hi,

Yes, of course. Wasn't that surprised actually, just eager to get an answer… You reckon this should be posted as a bug?

/Fredrik
 Report
User Avatar
aiworks
Member
130 posts
Joined: 4月 2009
Offline
 
Hi Frederik,

Might be a good idea. I'm actually surprised that none of the devs following the forums has not stumbled upon our discussion, yet.


Greg.
 Report
  • Quick Links