Prim Sorting

[sibarrick]

Small otl offering....

All it does is sort prims using a user entered order.

Don't know why the sort sop doesn't allow this.... but I got fed up waiting for the rfe to be implimented

OPsort.zip

[Mario Marengo]

Hahaaaaaa! That's Great!!!

"How to Twist and Bend Houdini To Do your Bidding", by Simon Barrick.

Very nice; thanks!

[sibarrick]

If there were more hours in a day than particles in the universe I could learn to program properly and use the HDK itself

Now, if they would just create a VOP style editor for HDK then sparks would really fly.

[edward]

I see. I guess the advantage for this method is that you can use normal patterns. Ordered groups might even work.

Otherwise, something like arg($PR, "2 3 4 1 0") would work in the Sort SOP.

[sibarrick]

Maybe it should work, but it doesn't. Not in H6.6.87 anyway.

I tried your expression in expressions and in shift and it don't do nothin'

Has this changed in H7 maybe? Please don't tell me I just wasted 10 minutes making this otl all for nowt.

[dobe]

Hello Simon,

reverse the order in the expression [arg(" 2 1 3 4 6 5 0 7 8" ,$PR) ] and it

works.

[sibarrick]

Ah yes, should have spotted that, since I use the arg expression in my otl.

Still doesn't do ranges though.

[anakin78z]

There's a problem with that. If you specify fewer arguments than faces, the result will not be correct.

Say you have a box, 6 faces. Take simons sop and specify '1 0 5 3 4'. It'll create the result you expect. Now try the sort sop with this expression: arg("1 0 5 3 4",$PR) . The result is different. If you do this: arg("1 0 5 3 4 2",$PR), it will work, but you had to give a number for every primitive. Say you have 13048 faces, and you want to order a few of those to number from 0 to 10? What then sesi, what then?

I was actually trying to find a nicer solution yesterday, cause I hate the delete sop, and I don't think Simons solution keeps point counts & numbers the same.

-z

[sibarrick]

How do you mean, doesn't keep point counts and numbers the same?

If i pump in an object with 90 points and x prims I get one out with 90 points and x prims.... or do you mean something else?

[edward]

Sorry, for the expression error. But you guys got my point anyhow.

Then you do something like this:

if( $PR < 5, arg("1 0 5 3 4", $PR), 2 )

?

[sibarrick]

nice, good for most situations, this should definitely be in the help.

Gets a bit hairy though if you need something complex, like 2-3 5 9-7 6 0 99-100 98-10 1......

I know you can do this, can't be bothered myself, maybe it could be an odchallenge!

[anakin78z]

How do you mean, doesn't keep point counts and numbers the same?

If i pump in an object with 90 points and x prims I get one out with 90 points and x prims.... or do you mean something else?

put down a box sop, append your Prim Sort sop, and enter a pattern. Whereas points were consolidated before, the geometry is now faceted, and you have 24 points instead of 8.

[sibarrick]

Oh i see what you mean i usually only use it for discreet polys. Surely you can just append a fuse sop to put it back how it started with just the 8 points?

[anakin78z]

sure, if you only started with 8 points!

[exel]

reverse the order in the expression [arg(" 2 1 3 4 6 5 0 7 8" ,$PR) ] and it

works.

Pretty nice, thanks Ed, Simon, etc...

So much of Houdini's power lies in little tricks like these that keep everything procedural and minimize "grunt work"...

I'd LOVE to see examples of these tricks start showing up in the SOP's help pages, thats the first thing everyone looks at when they've got a question, no?

-JS-

[sibarrick]

Sorry to drag this up again, but I've just realised that using the arg("3 2 1 0",$PR) method works completely oppositely to my primsort.otl.

So watch out. Basically in my method you enter the order that you want to the prims to be sorted into, whereas using the arg expression in a sort sop you do the reverse, you have to specify what prim number you want each current prim to turn into, in the order 0 1 2 3 4.. etc

Anyway it's difficult to describe but if you are interested. I can understand why you wouldn't be see the attached.

[sibarrick]

also, Anakin78z, I take your point. The original reason for this post really was to prove that this was a function that was required in Houdini. I think your point probably proves even more that it should be in there. Expressions and Otls can only go so far. It seems like such a basic function, and must be almost trivial to implement I don't know why it is missing.