Reduce points in Point feature from a 30 meter DEM raster

[ChrisMartin9]

I have created a point feature from a 30 meter DEM raster and I would like to reduce the number of points (decimate) - the number of points are too large to manage. Is there a toolbox for this or can this be done with the select attribute tool. Thanks.

[DanPatterson_Retired]

probably many options, but you could add a field to the table and calculate a random number for the field , then select the number of features you want.  If you want to do a rondom-y stratified-ish sample, you could calculate add a field and use the field calculation on either the X and/or Y

!shape!.centroid.X % 2     ( I think)  which will calculate a binary value indicating whether the coordinate is even or odd...then query out your choise and save them.  So decide whether you need random, random with a minimum number of samples, points with a minimum spacing etc etc.

Field calculations are here if you need examples.  Calculate Field examples—Help | ArcGIS for Desktop

[DuncanHornby]

I don't think this is really a Python question, more of a geo-processing problem (python being one solution).

Assuming you have a Shapefile then your row id is "FID". The following Where clause can be used in the Select by Attribute and it will select the first and every tenth row:

"FID" / 10 = round( "FID"/10,0) or "FID" = 1

If your dataset is a geodatabase feature class then substitute FID with ObjectID.

You can then either delete these or invert the selection and delete out the batches of 9 rows. If the second option is what you want it will be quicker to simply export the selected 10th rows.

[DanPatterson_Retired]

Duncan, I didn't mean to suggest it was a python problem, however, python in the field calculator etc allows for builtin random module access making random number generation easier etc.

[DuncanHornby]

Dan, I was not thinking your answer was unsuitable more that the original question is located under the python discussion board and that it is seems more suited to the general geo-processing page. If I had the power I would be moving it into that section.

[ChrisMartin9]

Duly noted but I was originally considering applying a Python script to execute this. I couldn't figure out the code to use in 'select by attribute' to delete rows in the table.

Thanks for your help.

[DuncanHornby]

Chris, If Dan's or my suggestion solves your problem you should tick it as such, shows to others a solution was found and we get lots of lovely points!

[ChrisMartin9]

Duncan,

I'm going to try this. I assume I need to start an edit session in order to delete the records in the attribute table?

[DuncanHornby]

You can or you can simply use the Delete Features geoprocessing tool! A warning about this tool, if no selection exists then it Deletes EVERYTHING and there is no undo for this tool! So if you are building a model to run this many times it would be sensible to put in a GetCount and use that as a precondition to the Delete Features tool, otherwise you may end up shooting yourself in the foot.

[ScottFraser3]

This was a great help.

1. "FID" / 10 = round( "FID"/10,0) or "FID" = 1 

I just adjusted the "10" counts until I got the spread I was seeking, which ended up being "6".

Sweet!