Python OGR: Filter Shapefiles by Lat Long and Export To CSV

Question

Using osgeo's Python API for OGR, I'd like to implement the following process:

1. Import ESRI Shapefiles from a given directory (call it input_dir)
2. Import a CSV file that contains lat-longs of interest (call it poicsv)
3. Join the CSV file long, lats with the Shapefile layer and extract the feature attribute values for that long, lat

I am currently up to the step where I have a single Shapefile loaded, and the CSV file loaded with long/lats in a list of tuples (ie. [(-123.76073, 41.24946), (-123.84451, 41.24946), ...]). Per long lat, I am trying to use SetSpatialFilter on the layer by the following code:

for ll in longlat:
    point = ogr.Geometry(ogr.wkbPoint)
    point.AddPoint_2D(ll[0], ll[1])
    layer.SetSpatialFilter(point)
    for feat in layer:
        feat.items()

However, upon sanity checking I've noticed there is no API for removing the Spatial Filter in my for loop, and I am also concerned about the memory leak created by creating a wkbPoint through the ogr.Geometry function in every iteration. Will the point object be cleaned up every iteration, and how could I reset the Spatial Filter every iteration?

Answer 1

Your point geom will get cleaned up automatically by Python's garbage collector each iteration after it goes out of scope. I just tested by running a 10 million iteration loop and creating a point and overwriting the same variable and memory usage stayed the same.

According to the documentation, you can reset your spatial filter by passing a NULL value (None in Python), i.e.

layer.SetSpatialFilter(None)

Though I don't think you actually need to as you're overwriting the existing spatial filter with a new one each iteration.