Mikhail Nikoliukin created SPARK-54718:
------------------------------------------
Summary: Preserve attributes names during CTE newInstance()
Key: SPARK-54718
URL: https://issues.apache.org/jira/browse/SPARK-54718
Project: Spark
Issue Type: Task
Components: SQL
Affects Versions: 4.2.0
Reporter: Mikhail Nikoliukin
Current implementation of CTERelationRef.newInstance() has incorrect behavior
if cte has multiple attributes with the same ID but different names. The common
case of it: referencing one attribute by names with different cases: {{SELECT
COLNAME, colname FROM t
}}
Together with the fact that newInstance() is called during exprId
deduplication, this leads to inconsistent results: sometimes names in one case,
sometimes in another
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
