[
https://issues.apache.org/jira/browse/SPARK-39510?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
zhengruifeng updated SPARK-39510:
---------------------------------
Description:
In Pandas-API-on-Spark:
1, *MonotonicallyIncreasingID* and *AttachDistributedSequence* are widely used
as the indexing type for big tables, they naturally have _RangePartitioning_
and _Ascending Ordering._
2, _Sorting by index_ is a common scenario, such as:
2.1 internally,
{color:#00875a}_Window.orderBy(NATURAL_ORDER_COLUMN_NAME)_{color} is widely
used in GroupBy, Rolling, Expanding, EWM, etc;
2.2 the {color:#00875a}_sort_index()_{color} should be a high
frequently used function for end users;{color}{color}
So we should make *MonotonicallyIncreasingID* and *AttachDistributedSequence*
provide the _RangePartitioning_ and _Ascending Ordering, so that the optimizer
can leverage them._
was:
In Pandas-API-on-Spark:
1, *MonotonicallyIncreasingID* and *AttachDistributedSequence* are widely used
as the indexing type for big tables, they naturally have _RangePartitioning_
and _Ascending Ordering._
2, _Sorting by index_ is a common scenario, such as:
2.1 internally,
{color:#00875a}_Window.orderBy(NATURAL_ORDER_COLUMN_NAME)_{color} is widely
used in GroupBy, Rolling, Expanding, EWM, etc;
2.2 the {color:#00875a}_sort_index()_ {color:#172b4d}should be a high
frequently used function for end users;{color}{color}
So we should make *MonotonicallyIncreasingID* and *AttachDistributedSequence*
provide the _RangePartitioning_ and _Ascending Ordering, so that the optimizer
can leverage them._
> leverage the natural partitioning and ordering of MonotonicallyIncreasingID
> ---------------------------------------------------------------------------
>
> Key: SPARK-39510
> URL: https://issues.apache.org/jira/browse/SPARK-39510
> Project: Spark
> Issue Type: Improvement
> Components: SQL
> Affects Versions: 3.4.0
> Reporter: zhengruifeng
> Priority: Major
>
> In Pandas-API-on-Spark:
> 1, *MonotonicallyIncreasingID* and *AttachDistributedSequence* are widely
> used as the indexing type for big tables, they naturally have
> _RangePartitioning_ and _Ascending Ordering._
> 2, _Sorting by index_ is a common scenario, such as:
> 2.1 internally,
> {color:#00875a}_Window.orderBy(NATURAL_ORDER_COLUMN_NAME)_{color} is widely
> used in GroupBy, Rolling, Expanding, EWM, etc;
> 2.2 the {color:#00875a}_sort_index()_{color} should be a high
> frequently used function for end users;{color}{color}
>
> So we should make *MonotonicallyIncreasingID* and *AttachDistributedSequence*
> provide the _RangePartitioning_ and _Ascending Ordering, so that the
> optimizer can leverage them._
--
This message was sent by Atlassian Jira
(v8.20.7#820007)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
