hive-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Brock Noland (JIRA)" <>
Subject [jira] [Commented] (HIVE-5264) SQL generated by not compliant with Postgres.
Date Wed, 25 Sep 2013 18:24:06 GMT


Brock Noland commented on HIVE-5264:

If you look all the way at the bottom of this:

you will see what I am referring to. The issue is at present that HiveQA cannot detect the
difference between a checkstyle failure and a compile failure since ant exists with 1 in both
cases. This means when trunk fails with checkstyle errors no tests will be run by Hive QA.
> SQL generated by not compliant with Postgres.
> ---------------------------------------------------------------------
>                 Key: HIVE-5264
>                 URL:
>             Project: Hive
>          Issue Type: Bug
>          Components: Metastore
>    Affects Versions: 0.12.0
>         Environment: Ubuntu 12.04
> PostgreSQL 9.1.8
>            Reporter: Alexander Behm
>            Assignee: Sergey Shelukhin
>         Attachments: D12993.1.patch, D12993.2.patch, HIVE-5264.01.patch, HIVE-5264.02.patch,
HIVE-5264.03.patch, HIVE-5264.patch
> Some operations against the Hive Metastore seem broken
> against Postgres.
> For example, when using HiveMetastoreClient.listPartitions()
> the Postgres logs show queries such as:
> 2013-09-09 19:10:01 PDT STATEMENT:  select PARTITIONS.PART_ID from
> join DBS on TBLS.DB_ID = DBS.DB_ID  where TBLS.TBL_NAME = $1 and
> DBS.NAME = $2 order by PART_NAME asc
> with a somewhat cryptic (but correct) error:
> ERROR:  relation "partitions" does not exist at character 32
> Postgres identifiers are somewhat unusual. Unquoted identifiers are interpreted as lower
case (there is no Postgres option to change this). Since the Metastore table schema uses upper
case table names, the correct SQL requires escaped identifiers to those tables, i.e.,
> select "PARTITIONS"."PART_ID" from "PARTITIONS"...
> Hive sets by default, so the above SQL is generated by
hive/metastore/, i.e., this is not a Datanucleus problem.
> When I set, then the Metastore backed by Postgres works.

This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see:

View raw message