[jira] [Commented] (SPARK-40137) Combines limits after projection

["Apache Spark (Jira)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/SPARK-40137?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17581248#comment-17581248 ] 

Apache Spark commented on SPARK-40137:
--------------------------------------

User 'ConeyLiu' has created a pull request for this issue:
https://github.com/apache/spark/pull/37565

> Combines limits after projection
> --------------------------------
>
>                 Key: SPARK-40137
>                 URL: https://issues.apache.org/jira/browse/SPARK-40137
>             Project: Spark
>          Issue Type: Improvement
>          Components: SQL
>    Affects Versions: 3.2.2
>            Reporter: Xianyang Liu
>            Priority: Major
>
> `Dataset.show` will add extra `Limit` and `Projection` on top of the given logical plan. If the `Dataset` is already a limit job, this will introduce an extra shuffle phase. So we should combine the limit after projection. 
> For example:
> ```scala
> spark.sql("select * from spark.store_sales limit 10").show()
> ```
> Before:
> ```
> == Physical Plan ==
> AdaptiveSparkPlan (12)
> +- == Final Plan ==
>    * Project (7)
>    +- * GlobalLimit (6)
>       +- ShuffleQueryStage (5), Statistics(sizeInBytes=185.6 KiB, rowCount=990)
>          +- Exchange (4)
>             +- * LocalLimit (3)
>                +- * ColumnarToRow (2)
>                   +- Scan parquet spark_catalog.spark.store_sales (1)
> +- == Initial Plan ==
>    Project (11)
>    +- GlobalLimit (10)
>       +- Exchange (9)
>          +- LocalLimit (8)
>             +- Scan parquet spark_catalog.spark.store_sales (1)
> ```
> After:
> ```
> == Physical Plan ==
> CollectLimit (4)
> +- * Project (3)
>    +- * ColumnarToRow (2)
>       +- Scan parquet spark_catalog.spark.store_sales (1)
> ```



--
This message was sent by Atlassian Jira
(v8.20.10#820010)

---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@spark.apache.org
For additional commands, e-mail: issues-help@spark.apache.org