How to delete double objects?

Hi all,
How can I delete twice or congruent objects in Rhino model space?

Replies to This Discussion

Permalink Reply by David Rutten on January 15, 2010 at 12:05pm

Hi Rafi,

what kind of objects? Points? Curves? BReps?

And, do you want to delete exact duplicates, or do you want to delete objects which kinda-sorta look alike... a bit... somewhat?

--
David Rutten
david@mcneel.com
Poprad, Slovakia

Permalink Reply by Suryansh Chandra on January 15, 2010 at 12:36pm

David,

Looking at the number of questions regarding this, perhaps it might be worthwhile having a component that does this? Perhaps it could even have a user controlled tolerance for the kinda-sorta look alike... a bit... somewhat cases?

Permalink Reply by David Rutten on January 15, 2010 at 2:22pm

Yes, that's what I figured. As soon as you allow tolerances it becomes a whole lot harder to solve these cases (except for points). Exact duplicates shouldn't be too hard.

--
David Rutten
david@mcneel.com
Poprad, Slovakia

Permalink Reply by Rafi on January 15, 2010 at 12:20pm

Hallo David,

I just want delete congruent curves and points which are exact duplicates!

Permalink Reply by David Rutten on January 15, 2010 at 2:27pm

Hi Rafi,

if you're looking at exact duplicates of curves and points, here's what I'd do:

1) Write a new class in a VB or C# component that implements IComparer(Of IOnCurve). You can check easy properties first (boundingbox, length, start and end points), if these all turn out to be the same, you'll eventually have to drill down all the way to Nurbs properties (i.e. compare the CV array and Knot-Vectors of the curves), but perhaps it's good enough to stick with the easy comparisons.

2) Iterate through a list of IOnCurves (or OnCurves, or OnNurbsCurves, whatever you've got), and slowly populate a SortedList or SortedDictionary as you loop. This SortedList must be aware of your Comparer. Then, whenever you encounter a curve which already has been handled before, you know you've got a duplicate.

This is not exactly trivial, not even for the somewhat experienced DotNetters. I'll see if I have some time tomorrow to implement this.

--
David Rutten
david@mcneel.com
Poprad, Slovakia

Permalink Reply by Suryansh Chandra on January 18, 2010 at 12:56pm

I was thinking this would have something to do with DataCRC (for exact matches)...?

Since you mention comparing the 'easy properties', does that imply you suggest using the IComparer for the kinda-sorta almost cases?

And if it were only limited to exact duplicates, which method would be faster -- IComparer or DataCRC?

Permalink Reply by Arthur Mamou-Mani on October 20, 2010 at 6:14am

Hi guys,

I have been using Kangaroo to delete duplicate points and so far it's been great but
when I increased the amount of points to 160,000, it crashes.

Would there be an alternative which could take such density or does it have to do with computer capacity ?

Many thanks,

Arthur

Permalink Reply by Shawn Fernandez on October 21, 2010 at 3:12pm

I just wrote this VB script component, not sure what the density capability is, but it removes duplicates for points, lines and curves. Let me know how it works for your large number of points.

Attachments:

Delete Duplicates.ghx, 220 KB

Permalink Reply by Rafi on January 15, 2010 at 1:34pm

Sorry!
I received just the 3D model from a customer, it is therefore not a Grasshopper Component to clean up!

Permalink Reply by Julian Loeffler on January 29, 2010 at 9:35am

I have exactly the same problem - In my case its identical points - or points that share the same z coordinate, but it would be great to have a component for all kinds of objects.

julian

Permalink Reply by Theunis du Toit on January 30, 2010 at 7:52am