[
https://issues.apache.org/jira/browse/KUDU-1872?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15898004#comment-15898004
]
Todd Lipcon commented on KUDU-1872:
-----------------------------------
In general, exposing partitioning metadata via an API would be useful for
Impala as well. For example, if two tables are joined on a column which is
being used for hash partitioning on both sides of the join, and the number of
hash partitions of one table is a multiple of the other, then a more efficient
join plan can be used.
> Expose range partitioning info in API
> -------------------------------------
>
> Key: KUDU-1872
> URL: https://issues.apache.org/jira/browse/KUDU-1872
> Project: Kudu
> Issue Type: Wish
> Components: client, impala
> Reporter: Przemyslaw Maciolek
>
> The API allows to create and drop range partitions using:
> [AlterTableOptions#addRangePartition
> (Java)|https://kudu.apache.org/apidocs/org/apache/kudu/client/AlterTableOptions.html#addRangePartition-org.apache.kudu.client.PartialRow-org.apache.kudu.client.PartialRow-]
> or [KuduTableAlterer#AddRangePartition
> (C++)|http://kudu.apache.org/releases/1.2.0/cpp-client-api/classkudu_1_1client_1_1KuduTableAlterer.html#a521a968f4db133a10629f76c75dde793]
> - however it is currently not possible to retrieve information about the
> currently available partitions. In effect, when partitions are to be added
> dynamically by the client, they either must be optimistically added (i.e.
> ignoring AlreadyExists error) or tracked separately.
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
