Re: bug report: slow getColumnTypeName - Mailing list pgsql-jdbc

From Luis Flores
Subject Re: bug report: slow getColumnTypeName
Date
Msg-id CA+nXnG97xdBY5RwZ7Pjz3fVE-9rakFJkSyGMoQOJB0EbUQYKRw@mail.gmail.com
Whole thread Raw
In response to Re: bug report: slow getColumnTypeName  (Maciek Sakrejda <m.sakrejda@gmail.com>)
List pgsql-jdbc
You're absolutely correct, I though the serial name was also reported
in psql, but no, only in pgadmin, so, like you said, in postgresql
there are only int4 and int8.

Which means we (or the driver maintainers, to be more precise) must
choose, postgresql complaint (int4) or jdbc complaint (serial)...

But changing this type of things is not nice for applications ...

Luis Flores


On Thu, Oct 11, 2012 at 4:45 PM, Maciek Sakrejda <m.sakrejda@gmail.com> wrote:
> On Thu, Oct 11, 2012 at 7:07 AM, Luis Flores <luiscamposflores@gmail.com> wrote:
>> If you use a column with a default value from a sequence, the column
>> type name doesn't change. If the column is int AND is generated from a
>> sequence called <TABLE>_<COLUMN>_seq, then in postgresql (you can
>> check in pgadmin) the column type name name changes from integer to
>> serial. We may or may not agree with this, but it's the way postgresql
>> works, so, the jdbc driver correctness and naming should be in line
>> with the postgresql server naming (mapping to jdbc type naming).
>
> For what it's worth, based on pg_attribute, this seems to be a
> pgAdmin-ism rather than the way Postgres works. The column type as
> reported by the catalogs is "integer" whether you create a serial-type
> column or an integer column with a default from your own sequence.


pgsql-jdbc by date:

Previous
From: Maciek Sakrejda
Date:
Subject: Re: bug report: slow getColumnTypeName
Next
From: Thor Michael Støre
Date:
Subject: ResultSetMetaData.isNullable(i) and outer joined columns