Re: Automating Partitions in PostgreSQL - Query on syntax

2009/4/21 Kedar Potdar <kedar(dot)potdar(at)gmail(dot)com>:
>
>
> 2009/4/21 Dickson S. Guedes <listas(at)guedesoft(dot)net>
>>
>> Em Ter, 2009-04-21 às 17:20 +0530, Kedar Potdar escreveu:
>> > Hi ,
>> >
>> > We are working on a patch to automate partitioning in PostgreSQL.
>>
>> Nice. ":)
>>
>> > For Range partitions, we have proposed the syntax which is as
>> > follows
>> >
>> > (...)
>> > PARTITION BY RANGE (emp_id)
>> > (
>> > emp_500 (START 1 END 500),
>> > emp_1500 (START 500 END 1500),
>> > emp_4000 (START 1520 END 4000)
>> > );
>>
>> What if I need more columns to set the partitions?
>
>
> You can do so by using command like this,
>
> CREATE TABLE emp (
>     emp_id          int not null primary key,
>     designation     text not null,
>     location        varchar(50) not null,
>     jdate           date not null,
>     ctc             float not null
>
> )
> PARTITION BY RANGE (jdate, emp_id)
> (
> emp_500 (START '01-01-1980',1 END '01-01-1990',500),
> emp_1500(START '01-01-1990',500 END '01-01-2000',1500+10),
> emp_4000 (START '1 Jan 2000',1500+20 END 'Jan 1, 2010', 4000-50)
> );
>
> You can have multiple columns as partition key attributes and values for
> these attributes should appear in the order specified.

-1, this is really ugly

Pavel

>
>>
>>
>>
>> []s
>> --
>> Dickson S. Guedes
>> mail/xmpp: guedes(at)guedesoft(dot)net - skype: guediz
>> http://guedesoft.net - http://planeta.postgresql.org.br
>
>