Python UDTF

data-grassroots — Tue, 01 Jul 2025 16:40:47 GMT

Attempting to create a Python UDF and receiving an error stating it's not supported in my environment.

running on SQL Serverless Cluster - Preview (v 2025.16)

[FEATURE UNAVAILABLE]

Python UDTF is not supported in your environment.

To use this feature, please contact Databricks Support. SQLSTATE: 56038

create function _rrr_taxonomy_test()

RETURNS TABLE(dept string, cat string, subcat string)

LANGUAGE PYTHON

s = '''select

custom_department as dept,

custom_category as cat,

coalesce(custom_sub_category, '') as subcat

from item'''

return spark.sql(s)

Re: Python UDTF

nayan_wylde — Tue, 01 Jul 2025 18:36:04 GMT

Currently Serverless supports only scaler valued function not table valued function. also you need to provide the catalog and schema in the function.

You can create using a pro compute

create function {catalog_name}.{schema_name}._rrr_taxonomy_test()

RETURNS TABLE(dept string, cat string, subcat string)

LANGUAGE PYTHON

s = '''select

custom_department as dept,

custom_category as cat,

coalesce(custom_sub_category, '') as subcat

from item'''

return spark.sql(s)

I tried this function in serverless and it worked it is a scaler valued function.

create function {catalog_name}.{schema_name}._rrr_taxonomy_test_1()

RETURNS string

LANGUAGE PYTHON

s = '"Hello World"

return s