Rtrees

Replies to This Discussion

Permalink Reply by Arend on July 6, 2016 at 9:33am

Perhaps this helps you? This is a simple example that finds points and their indices. (Example included for a set of 10.000 points)

  private void RunScript(List pointList, Point3d startPoint, double radius, ref object A, ref object B)

  {

    try {

      RTree tree = new RTree();

// first insert all points.
      for (int i = 0; i < pointList.Count; i++)

      {

        // add a point, and a number, as a reference for this point.

        // this can also be a boundingbox.

        tree.Insert(pointList[i], i);

      }

// we'll add our matches to this list.
      List closestPoints = new List ();

      List closestIndices = new List ();

tree.Search(new Sphere(startPoint, radius), (object sender, RTreeEventArgs events) => {
        // this will be executed for each point that matches the search criteria

        RTreeEventArgs e = events;

// look up which point this is.
        closestPoints.Add(pointList[e.Id]);

        closestIndices.Add(e.Id);

        });

A = closestPoints;
      B = closestIndices;

    } catch (Exception e)

    {

      throw new Exception(e.Message + " " + e.StackTrace);

    }

  }

Attachments:

• example-rtree.gh, 269 KB

Permalink Reply by Arend on July 7, 2016 at 9:58am

Interesting. This are the headaches of optimisation - which optimisation technique works best for which problem?

I can't check right now, but in most cases if it's the first time you execute a C# block it's the compilation of the block that takes the most time. Also - most of the time when I use RTree it's for the bounding box cases (might two objects possibly intersect, or not?), it might very well be that the overhead of building of a rtree for a simple one to many proximity search is always higher than a brute force method.

Permalink Reply by Jon on April 15, 2020 at 9:45am

Hi Arend,

Thanks for your above example.  I am also new to C#, and your code has been very useful.

I was wondering, what would be the simplest way to modify this code in order to use it with a list of points rather then a single point.

So, for each point in ListA, find the closest point in List B(within specified radius).
OR, if i can find for each point in ListA, the closest points in ListB(within specified radius) in order of closest distance.  

i know i need to search for each point of ListA (iMeshPoints)... but then i am lost...

for (int i = 0; i < iMeshPoints.Count; i++)
{}

Hope it hasn't been too long since you've posted... Thanks! 

Permalink Reply by Arend on April 20, 2020 at 8:30am

Hi Jon,

I'd be happy to answer this question, but would you mind re-posting your question in the current forum at https://discourse.mcneel.com ? Then it's easier to answer, and also accessible for all other users.