[
https://issues.apache.org/jira/browse/KYLIN-1702?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15413083#comment-15413083
]
Zhong Yanghong commented on KYLIN-1702:
---------------------------------------
There's an issue when we do cube building at the step of building dictionary.
It cost too much time to check the duplicated contents of snapshot table,
especially when the snapshot is too large and the contents are stored at HDFS
rather than at the hbase.
> The Key of the Snapshot to the related lookup table may be not informative
> --------------------------------------------------------------------------
>
> Key: KYLIN-1702
> URL: https://issues.apache.org/jira/browse/KYLIN-1702
> Project: Kylin
> Issue Type: Improvement
> Reporter: Zhong Yanghong
> Assignee: Zhong Yanghong
>
> Currently the key for the snapshot stored in hbase metadata file is as
> follows:
> ResourceStore.SNAPSHOT_RESOURCE_ROOT + "/" + new
> File(signature.getPath()).getName() + "/" + uuid + ".snapshot"
> However, some hive tables stored in hive may organized like
> dirName/tableName/00, dirName/tableName/01.
> Based on current setting, the key will be
> ResourceStore.SNAPSHOT_RESOURCE_ROOT + "/" + 00 + "/" + uuid + ".snapshot",
> which is lack of the table name information.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)