Python API reference

The Python API reference documents Sedona's Python modules, classes, and functions, including their parameters and return types.

For functions used with PySpark DataFrames, see the SQL function reference. For example, ST_Length calculates the length of a linestring geometry:

from sedona.spark.sql.st_functions import ST_Length

df.select(ST_Length("geometry").alias("length"))

This example assumes a DataFrame named df with a geometry column named geometry and a Spark session configured with Sedona. See the Python installation guide for setup instructions and the DataFrame API guide for argument conventions and more examples.

Recognizing spatial column types

Since v2.0.0, is_geometry_type and is_geography_type recognize both Sedona UDTs and Spark native spatial types when those classes are available in PySpark. They inspect schema type instances without requiring a Spark session or changing the column's representation.

Both helpers are available from sedona.spark, sedona.spark.sql and sedona.spark.sql.types.

from sedona.spark import is_geography_type, is_geometry_type

geometry_columns = [
    field.name for field in df.schema.fields if is_geometry_type(field.dataType)
]
geography_columns = [
    field.name for field in df.schema.fields if is_geography_type(field.dataType)
]

The predicates distinguish geometry from geography. Arrays and structs containing spatial values are not themselves spatial column types. LegacyGeometryType and LegacyGeographyType in sedona.spark.sql.types explicitly name Sedona's UDTs; the predicates recognize these too. These helpers do not enable native spatial support or change the existing GeometryType and GeographyType defaults.