Thread: Patch for automated partitioning

Patch for automated partitioning

From
shrish purohit
Date:
Hi,

Updated Kedar's patch sending on behalf of Kedar.

.tar.bz2 file contains files to be put at
./src/include/catalog/pg_partition.h
./src/backend/utils/partition
./src/test/regress/sql/partition.sql

There are not toomany modifications. It includes improvements for validating multi-column partitioned tables.
propagation of constraints and indexes to child tables. addition of minvalue and maxvalue. some bug fixes.

Thanks,
Shrish
Persistent Systems Ltd.
www.persistentsys.com
Attachment

Re: Patch for automated partitioning

From
Peter Eisentraut
Date:
On Fri, 2009-10-30 at 17:55 +0530, shrish purohit wrote:
> Hi, 
> 
> Updated Kedar's patch sending on behalf of Kedar.
> 
> .tar.bz2 file contains files to be put at 
> ./src/include/catalog/pg_partition.h
> ./src/backend/utils/partition
> ./src/test/regress/sql/partition.sql
> 
> There are not toomany modifications. It includes improvements for
> validating multi-column partitioned tables.
> propagation of constraints and indexes to child tables. addition of
> minvalue and maxvalue. some bug fixes.
> 

Thank you for your contribution, but please read
<http://petereisentraut.blogspot.com/2009/09/how-to-submit-patch-by-email.html> on how to submit a proper patch by
email. A tarball of scattered files is not a convenient way of working, especially for a nontrivial feature like this.
 




Re: Patch for automated partitioning

From
Itagaki Takahiro
Date:
shrish purohit <shrishpurohit@gmail.com> wrote:

> Updated Kedar's patch sending on behalf of Kedar.
> There are not toomany modifications. It includes improvements for validating
> multi-column partitioned tables.

Thanks, but I'd like to keep the first patch as possible as simple.
I'll refer your work in the *next* step, especially partition
triggers and ALTER commands.

I glanced over your patch and find some codes that have room for improvement:
 * Need to avoid RANGE and PARTITION to be reserved keywords. * Non SQL-based operations are preferred if possible, so
needto   reconsider pg_parse_query(). * Adding NOTICE message to completionTag is a very bad idea.   we don't need such
featurebecause users can create trigger   on the overflow partition to send those messages.
 

Regards,
---
ITAGAKI Takahiro
NTT Open Source Software Center