Hyukjin Kwon created SPARK-61063:
------------------------------------
Summary: Take NumPy arrays, ints for dates and nested duplicate
field names in createDataFrame of local rows in the Rust client as PySpark does
Key: SPARK-61063
URL: https://issues.apache.org/jira/browse/SPARK-61063
Project: Spark
Issue Type: Bug
Components: Connect
Affects Versions: 5.0.0
Reporter: Hyukjin Kwon
createDataFrame of local rows differs from PySpark 4.3.0: a NumPy array in an
array field raises AssertionError (PySpark takes it when the elements need no
converter), an int for a DateType field raises ValueError (PySpark takes it as
days since the epoch), and a nested struct with duplicate field names sends IPC
that the server cannot read (PySpark deduplicates the Arrow names).
--
This message was sent by Atlassian Jira
(v10.3.26#10030026)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
