[
https://issues.apache.org/jira/browse/SPARK-59140?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Hyukjin Kwon reassigned SPARK-59140:
------------------------------------
Assignee: Hyukjin Kwon
> Close PySpark 4.2.0 public-API parity gaps in the Rust drop-in client
> ---------------------------------------------------------------------
>
> Key: SPARK-59140
> URL: https://issues.apache.org/jira/browse/SPARK-59140
> Project: Spark
> Issue Type: Improvement
> Components: Connect
> Affects Versions: 5.0.0
> Reporter: Hyukjin Kwon
> Assignee: Hyukjin Kwon
> Priority: Major
> Labels: pull-request-available
> Fix For: connect-rust-4.2.0
>
>
> The Rust drop-in client (pyspark-client-rust) advertises "100% public-API
> parity with PySpark 4.2.0 / runs unchanged", but keyword-argument calls
> raised TypeError on ~507 SQL functions and dozens of
> DataFrame/Column/Catalog/Reader/Writer/Streaming/Session methods, and several
> parameters were missing entirely. So common PySpark code did not run
> unchanged.
> This issue tracks bringing the public API of pyspark.sql.connect.* to
> keyword-argument and signature parity with PySpark 4.2.0.
> Verified with an independent signature harness against the v4.2.0 tag:
> - function keyword-arg mismatches: 513 -> 0
> - class-method param mismatches: 70 -> 2 (both intentional: Column.alias/name
> expose an explicit metadata= where the reference uses **kwargs)
> Also verified: all Rust golden/core tests pass, and a focused end-to-end
> suite runs 16/16 green against a real Spark Connect 4.2.0 server.
> A few params are accepted for signature parity but not yet behaviorally wired
> pending connect-core/proto support (documented in-code and in
> PARITY_GAP_ANALYSIS.md): mapInPandas/mapInArrow(profile),
> localCheckpoint(storageLevel), printSchema(level),
> createDataFrame(verifySchema), trigger(realTime), multi-path
> read.parquet(*paths).
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
