The usual way to use Apache Sedona inside pySpark is by first registering Sedona types and functions with

SedonaRegistrator.registerAll(spark)

We need to have these autoregistered when the cluster start (to be able, for example, to perform geospatial queries with the Databricks SQL Connector for Python).

From my understanding, autoregistration can be obtained by adding the following cluster configuration, but it doesn't work:

spark.sql.extensions org.apache.sedona.sql.SedonaSqlExtensions

Do I miss something?

Is my expectation wrong?