I have a multi point Townland file. I want to be able to remove duplicates within a radius of 3000 meters from each individual point??

I cannot simply remove duplicates as they are from various locations.

E.g. There may be five points called Castletown within a 4000m radius and I only want one.

However there may be another five points called Castletown 40,000m away (As more than one area is called Castletown) and I also only want to keep one of these within a 4000m radius of itself.

Therefore I would end up with two points called Castletown from two separate locations.

Any ideas whats the best way to do this?

2 Answers 2


You can run the Delete Identical tool using either the shape geometry (if the geometries are identical) or a named field, such as "Casteltown". This will remove all but one of the points. You can test this by using the Find Identical tool, which won't delete the data.

The Delete Identical tool also has a tolerance setting, which will allow you to filter by location. This tool is your best bet if you'd like to process an entire point file at once using a standard search tolerance (e.g. the tolerance of 4000m can be applied to all points since no two "Castletowns" are within 4000m of each other).

Note: This tool requires ArcInfo.


You could approach this using a Python script. You would have to iterate over all of the point features and create a 3000m buffer for each one. If there are multiple Castletowns within that buffer you have a few options to determine which one to keep:

  1. Keep the original one.
  2. Use the Central Feature tool which "Identifies the most centrally located feature in a point, line, or polygon feature class."
  3. Use the Mean Center tool which "Identifies the geographic center (or the center of concentration) for a set of features."
  4. Use the Median Center tool which "Identifies the location that minimizes overall Euclidean distance to the features in a dataset."

The tools listed above are available for all license levels of ArcGIS.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.