The first time I call to_datetime() in a Databricks notebook, it throws the warning: In Python 3.6+ and Spark 3.0+, it is preferred to specify type hints for pandas UDF instead of specifying pandas UDF type, which will be deprecated in the future releases. See SPARK-28264 for more details.
But this may be an incorrect warning, since to_datetime() does not seem to relate to SPARK-28264.
The warning does not happen when the same code is run again in the same notebook session.
The first time I call to_datetime() in a Databricks notebook, it throws the warning:
In Python 3.6+ and Spark 3.0+, it is preferred to specify type hints for pandas UDF instead of specifying pandas UDF type, which will be deprecated in the future releases. See SPARK-28264 for more details.
But this may be an incorrect warning, since to_datetime() does not seem to relate to SPARK-28264.
The warning does not happen when the same code is run again in the same notebook session.
This issue is partly addressed in https://github.com/databricks/koalas/issues/2044, but is not really answered there.