You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@trafodion.apache.org by "David Wayne Birdsall (JIRA)" <ji...@apache.org> on 2016/06/02 16:38:59 UTC
[jira] [Commented] (TRAFODION-1568) Need to create a partitioned
salted table with the CREATE TABLE LIKE DDL even though the source table is
not partitioned.
[ https://issues.apache.org/jira/browse/TRAFODION-1568?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15312619#comment-15312619 ]
David Wayne Birdsall commented on TRAFODION-1568:
-------------------------------------------------
Hive tables don't have a primary key. Salting requires a primary key. One possibility is to use a clustering key; perhaps salting can then be used.
> Need to create a partitioned salted table with the CREATE TABLE LIKE DDL even though the source table is not partitioned.
> ---------------------------------------------------------------------------------------------------------------------------
>
> Key: TRAFODION-1568
> URL: https://issues.apache.org/jira/browse/TRAFODION-1568
> Project: Apache Trafodion
> Issue Type: Bug
> Components: sql-cmu
> Reporter: Qifan Chen
> Assignee: David Wayne Birdsall
>
> When update stats for a native Hive table, the native table can be non-partitioned as far as Trafodion is concerned. Due to the data size of the sampling, it is important to have the capability to create a partitioned salted table with the CREATE TABLE LIKE cause even though the source table is not partitioned, with the syntax similar to the following.
> CREATE TABLE TRAFODION."_HIVESTATS_".SAMPLE_TABLE
> LIKE TRAFODION."_HV_HIVE_".CUSTOMER
> salt using 5 partitions on( C_CUSTOMER_SK,C_CUSTOMER_ID );
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)