Hyukjin Kwon created SPARK-61039:
------------------------------------
Summary: Send the eager and storageLevel of checkpoint and
localCheckpoint in the Python client as PySpark does
Key: SPARK-61039
URL: https://issues.apache.org/jira/browse/SPARK-61039
Project: Spark
Issue Type: Bug
Components: Connect
Affects Versions: connect-rust-4.3.0
Reporter: Hyukjin Kwon
In PySpark Connect, {{df.checkpoint(eager)}} and {{df.localCheckpoint(eager,
storageLevel)}} send {{eager}}, and for a local checkpoint the
{{storageLevel}}, in their {{CheckpointCommand}}. A checkpoint that is not
eager is not materialized until it is used, and a local one is kept at the
storage level given.
The Rust-backed Python client, and the Rust API, ignore them: the checkpoints
are always eager, at the default storage level. Send them as PySpark does.
--
This message was sent by Atlassian Jira
(v10.3.26#10030026)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
