Hyukjin Kwon created SPARK-60955:
------------------------------------
Summary: Make pyspark.ml.functions vector_to_array and
array_to_vector work in Spark Connect Rust
Key: SPARK-60955
URL: https://issues.apache.org/jira/browse/SPARK-60955
Project: Spark
Issue Type: Bug
Components: Connect
Affects Versions: connect-rust-4.3.0
Reporter: Hyukjin Kwon
In PySpark Connect, {{pyspark.ml.functions.vector_to_array}} and
{{array_to_vector}} build their columns through
{{pyspark.ml.connect.functions}}. The Rust-backed Python client has no
{{pyspark.ml.connect.functions}}, so every call raises {{ImportError: cannot
import name 'functions' from 'pyspark.ml.connect'}}, where {{spark.sql("SELECT
array(1.0D, 2.0D) a").select(vector_to_array(array_to_vector("a")))}} gives
{{[1.0, 2.0]}} in PySpark 4.3.0. Provide the module so that both functions work
as in PySpark.
--
This message was sent by Atlassian Jira
(v10.3.26#10030026)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
