Home | About | Sematext search-lucene.com search-hadoop.com
 Search Hadoop and all its subprojects:

Switch to Threaded View
Hive >> mail # user >> partition column not allowed in clustered by clause ?


Copy link to this message
-
Re: partition column not allowed in clustered by clause ?
Hi Ameet,
That's the correct behaviour.

In Hive, clustering and sorting happens within a partition. Inside each partition, there is only one value associated with the partition column therefore it would have no impact on clustering and sorting. Therefore, putting the partition column in clustered by/sorted by in the create table statement doesn't make sense.

Also, your create table statement should be something like (note the removal of col3 from the column list):
create table abc ( col1 string, col2 string)
partitioned by (col3 string)
clustered by (col1) sorted by (col1) into 10 buckets;

Mark

Mark Grover, Business Intelligence Analyst
OANDA Corporation

www: oanda.com www: fxtrade.com

----- Original Message -----
From: "ameet chaubal" <[EMAIL PROTECTED]>
To: [EMAIL PROTECTED]
Sent: Thursday, May 10, 2012 2:24:02 PM
Subject: partition column not allowed in clustered by clause ?

Hi All,
I am not able to create a table with partition column also included in the clustered by clause.
create table abc ( col1, col2, col3 )
partitioned by (col3)
clustered by (col1,col3) sorted by (col1,col3) into 10 buckets;
fails with : FAILED: Error in semantic analysis: Invalid column reference
Any reason why this is the case?

Sincerely,
Ameet