[
https://issues.apache.org/jira/browse/SPARK-4886?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14253154#comment-14253154
]
Xudong Zheng commented on SPARK-4886:
-------------------------------------
Hi Guowei,
"CACHE TABLE ... AS SELECT..." will create a new cache table instead of caching
the partition of the original table. The query on original table will still go
to HDFS. And this is not convenient for appending scenario, because will need
to create a new table every time we add a new partition. Actually, that is
still table level cache control.
> Support cache control for each partition of a Hive partitioned table
> --------------------------------------------------------------------
>
> Key: SPARK-4886
> URL: https://issues.apache.org/jira/browse/SPARK-4886
> Project: Spark
> Issue Type: Improvement
> Components: SQL
> Reporter: Xudong Zheng
>
> SparkSQL currently don't support control cache for each partition of a Hive
> partitioned table. If we could add this feature, user could have a better
> cache control of a cache table. And in many scenarios, the data is
> periodically appended into a table as a new partition, with this feature,
> user could easily control a sliding windows of data to be cached in memory.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
