Selecting only one point in polygon using QGIS

Question

I have two shapefiles: polygon(buffer) and point.

How can I select only one point in each buffer in QGIS?

If I use the "Select by Location", I will get all points inside the polygons, however I need only one? It doesn't matter which point out of two or more in polygon is selected.

Answer 1

If you don't care which point is selected, you can use this small PyQGIS script. Paste into a blank editor in the Python Console, change the layer names in the first two lines to match your layer names and click run.

# Change layer names in the first two lines to match your layers...
point_lyr = QgsProject.instance().mapLayersByName('New scratch layer')[0]
poly_lyr = QgsProject.instance().mapLayersByName('Buffered')[0]

ids = []

for feat in poly_lyr.getFeatures():
    contained_point = [f for f in point_lyr.getFeatures() if f.geometry().within(feat.geometry())][0]
    ids.append(contained_point.id())

point_lyr.selectByIds(ids)

Result:

Answer 2

With the "Select by expression" tool you can use:

array_first(
    array_agg(
        $id,
        group_by:=overlay_intersects('polygons', $id), -- replace 'polygons'
        filter:=array_length(overlay_intersects('polygons', $id)) > 0 -- replace 'polygons'
        )
    ) = $id

on your point layer. Replace the name of the 'polygons' layer with the actual name and run it.

QGIS 3.16+ is required.

Answer 3

One way to do it would be:

• First do a "Join Attributes by location" (Vector -> Data Management Tools) with your points as the Base Layer, your polygons as the Join Layer, "within" as the Geometric predicate and for the Join type : "Take attributes of the first matching feature only (one-to-one)". Remember you need to have a Spatial index for both your layers before doing this (Vector -> Data Management Tools -> Create Spatial Index)
• Then, now that all your points within the same polygon have the same attribute, use the "Delete Duplicates by attribute" from the Processing toolbox in order to keep only one point inside each polygon.

Hope that helps.

Answer 4

To add some performance improvements to @Ben W's answer by using a spatial index and a break command:

point_lyr = QgsProject.instance().mapLayersByName('points')[0]
poly_lyr = QgsProject.instance().mapLayersByName('polygons')[0]

ids = []
point_lyr.removeSelection() #clear selection

point_lyr_idx = QgsSpatialIndex(point_lyr.getFeatures()) # build a spatial index on the points

for polygon in poly_lyr.getFeatures():
    for pointid in point_lyr_idx.intersects(polygon.geometry().boundingBox()): # iterate only over points inside bbox of polygon
        point = point_lyr.getFeature(pointid)
        if point.geometry().within(polygon.geometry()): # check if point is actually inside polygon
            ids.append(point.id())
            break # stop iteration after first match and go to next polygon

point_lyr.selectByIds(ids)

Answer 5

To add even more performance improvements to @MrXsquared's answer by using QgsGeometryEngine and skipping feature fetching:

point_lyr = QgsProject.instance().mapLayersByName('points')[0]
poly_lyr = QgsProject.instance().mapLayersByName('polygons')[0]

ids = []
point_lyr.removeSelection() #clear selection

# build a spatial index on the points (with flag to be able to retrieve geometries)
point_lyr_idx = QgsSpatialIndex(point_lyr.getFeatures(), flags=QgsSpatialIndex.FlagStoreFeatureGeometries) 

for polygon in poly_lyr.getFeatures():
    poly_geometry_engine = QgsGeometry.createGeometryEngine(polygon.geometry().constGet())
    poly_geometry_engine.prepareGeometry()

    for pointid in point_lyr_idx.intersects(polygon.geometry().boundingBox()): 
        # iterate only over points inside bbox of polygon
        point_geometry = point_lyr_idx.geometry(pointid)

        if poly_geometry_engine.contains(point_geometry.constGet()): # check if point is actually inside polygon
            ids.append(pointid)
            break # stop iteration after first match and go to next polygon

point_lyr.selectByIds(ids)

QgsGeometryEngine can be really powerful in some cases so it is useful to know about it (in some cases I have gotten a x100 performance improvement using it).

With these improvements you can get a ~20-30% improvement compared to @MrXsquared's answer.

Answer 6

You don't need to write a script, there is a built-in Random selection within subsets tool. Before start you need to join by location ID of the polygons to your point data and specify that field in the ID field parameter.