Hi,
I would like to know how i could change (delete) these normals without altering the other numbers of the others normals, the numbers of the vertex and the other procedures of the network.

Greetings..
Navarro

You can't as those are inline points ….

It looks like what you've got there are two degenerate polygons. The polygon numbers are 4 and 65, so to delete the normals, you have to delete the polygons themselves (which I think is what you're asking)?

Unfrotunately, deleting those primitives will renumber the primitives. One guess is primitives 0, 1, 2 and 3 will remain intact, but everything until 65 will shift down one, and everything after 64 will shift down two. In Houdini, you can't have a gap in the sequence of primitive numbers.

Hope that helps,
George.

if only we had id's……

Hi,
Thanks George. The problem comes from an incorrect blast. then, i have 2 possible solutions:
1. -Replace the blast for a correct blast.

Problem: houdini changes the other vertex number and the edge number, change the geometry.

2. -Delete the normals at the end of the network.

Problem: Renumber the normals, change the geometry.

is there any way to correct the SOPs process (blast, polyextrude, polysplit,..) in the network

without modify the geometry parameters (vertex, edges and normal numbers)?

Gretings..
Navarro.

is there any way to correct the SOPs process (blast, polyextrude, polysplit,..) in the network

If you delete the offending polygons at the end of your chain, then yes, your primitives will get renumbered, but no harm should actually come to your model (since the changes are happening upstream).

If you delete the polygons up in the chain, then likely the rest of your operations will need tweaking (to say the least). You could theoretically go in and renumber each SOP after the one you do the deletion in. (In each SOP there should be a Group field that you can edit). But that sounds like such a waste of time. If you can the best thing to do is clean up your geometry the way you like and delete the history and continue modelling.

Like Simon said, we would need something like id's to prevent this sort of problem. Id's or some sort of spatial grouping, or voodoo magic…

Take care,
George.