Hyukjin Kwon created SPARK-61221:
------------------------------------
Summary: createDataFrame of local rows panics or raises a raw
Arrow error for a non-nullable NullType
Key: SPARK-61221
URL: https://issues.apache.org/jira/browse/SPARK-61221
Project: Spark
Issue Type: Bug
Components: Connect
Affects Versions: connect-rust-4.3.0
Reporter: Hyukjin Kwon
Repro: spark.createDataFrame([Row(a=[None])], StructType([StructField("a",
ArrayType(NullType(), False))])), which panics. Official PySpark 4.3.0 raises
ValueError "A null type field may not be non-nullable" for it and for every
such schema with non-empty data, from its Arrow schema conversion, as PyArrow's
pa.field rejects the field.
The fix raises that ValueError for non-empty local rows once their values are
converted, as PySpark does, and makes the Arrow list build of the Rust core
return an error instead of panicking.
--
This message was sent by Atlassian Jira
(v10.3.26#10030026)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
