Hyukjin Kwon created SPARK-61065:
------------------------------------
Summary: Export the names of PySpark from pyspark.sql.functions,
and type the literals of NumPy scalars, in the Rust client as PySpark does
Key: SPARK-61065
URL: https://issues.apache.org/jira/browse/SPARK-61065
Project: Spark
Issue Type: Bug
Components: Connect
Affects Versions: 5.0.0
Reporter: Hyukjin Kwon
`from pyspark.sql.functions import *` exports 18 names that PySpark 4.3.0 does
not, which shadow the builtin chr and the stdlib modules random and uuid, and
pyspark.sql.functions has from_avro and to_avro, which PySpark does not (they
are in pyspark.sql.avro.functions). lit() of NumPy scalars types them wrongly:
np.int64 as int instead of bigint, np.float32 as double instead of float, and
unsigned ints as int where PySpark raises UNSUPPORTED_DATA_TYPE.
--
This message was sent by Atlassian Jira
(v10.3.26#10030026)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
