[
https://issues.apache.org/jira/browse/SPARK-39169?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Vitalii Li updated SPARK-39169:
-------------------------------
Description:
When `FIRST` is a single aggregate function in `Aggregate` we could either
rewrite whole query or optimize execution logic.
* Plan => `SELECT FIRST(<col>) FROM <table>` => `SELECT <col> FROM <table>
LIMIT 1`. Note that setting `ignoreNulls` to `true` should block such rewrite
since returns could differ in case all values of <col> are `NULL`
* Execution => `SELECT FIRST(<col>) FROM <table> GROUP BY <some_col>` => short
circuit iteration per key once a value for `FIRST` is set.
was:
When `FIRST` is a single aggregate function in `Aggregate` we could either
rewrite whole query or optimize execution logic.
* Plan => `SELECT FIRST(<col>) FROM <table> [GROUP BY <col>]` => `SELECT <col>
FROM <table> LIMIT 1`. Note that setting `ignoreNulls` to `true` should block
such rewrite since returns could differ in case all values of <col> are `NULL`
* Execution => `SELECT FIRST(<col>) FROM <table> GROUP BY <col2>` => short
circuit iteration per key once a value for `FIRST` is set.
> Optimize FIRST when used as non-aggregate
> -----------------------------------------
>
> Key: SPARK-39169
> URL: https://issues.apache.org/jira/browse/SPARK-39169
> Project: Spark
> Issue Type: Improvement
> Components: SQL
> Affects Versions: 3.3.0
> Reporter: Vitalii Li
> Priority: Major
>
> When `FIRST` is a single aggregate function in `Aggregate` we could either
> rewrite whole query or optimize execution logic.
> * Plan => `SELECT FIRST(<col>) FROM <table>` => `SELECT <col> FROM <table>
> LIMIT 1`. Note that setting `ignoreNulls` to `true` should block such rewrite
> since returns could differ in case all values of <col> are `NULL`
> * Execution => `SELECT FIRST(<col>) FROM <table> GROUP BY <some_col>` =>
> short circuit iteration per key once a value for `FIRST` is set.
--
This message was sent by Atlassian Jira
(v8.20.7#820007)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
