I use pandas_udf to calculate intersection of points and polygons. When i use sample data frames to calc intersection with GEOPandas, not using pySpark, i get one intersection. But, then i use pyspark i get 2 intersection. Why it is happen? The point is very close to the border, but not in the border. Type of coordinates of point i use double. EPSG:4326, pyarrow = 0.14.1, geopandas = 0.9.0, pyspark = 2.3.1