I think it didn't reach the mail group. so i am sending it again

Jaison Joseph wrote:

>
>
> On Thu, Aug 13, 2009 at 11:42 AM, Dmitry Yemanov
> <dimitr@... <mailto:dimitr@...>>
> wrote:
>
>
> How wide are records of trg_trg?
>
>
> Here is the full ddl for trg_trg
>
> CREATE TABLE TRG_TRG
> (
> TRG_ID Integer,
> TRG_EMP_ID Integer,
> TRG_DTTM Timestamp,
> TRG_CODE Integer,
> TRG_PRSD Smallint DEFAULT 0,
> TRG_TRD_ID Integer,
> TRG_DOR_CODE Varchar(16),
> TRG_DEV_ID Integer,
> CONSTRAINT PK_TRG_TRG PRIMARY KEY (TRG_ID)
> );
>
> ALTER TABLE TRG_TRG ADD CONSTRAINT FK_TRG_TRG_EMP_EMP
> FOREIGN KEY (TRG_EMP_ID) REFERENCES EMP_EMP (EMP_ID) ON DELETE CASCADE;
> ALTER TABLE TRG_TRG ADD CONSTRAINT FK_TRG_TRG_TRG_DEV
> FOREIGN KEY (TRG_DEV_ID) REFERENCES TRG_DEV (DEV_ID);
> ALTER TABLE TRG_TRG ADD CONSTRAINT FK_TRG_TRG_TRG_TRD
> FOREIGN KEY (TRG_TRD_ID) REFERENCES TRG_TRD (TRD_ID);
> CREATE INDEX IX_TRG_TRG_DTTM ON TRG_TRG (TRG_DTTM);
> CREATE DESCENDING INDEX IX_TRG_TRG_DTTM_DESC ON TRG_TRG (TRG_DTTM);
> CREATE UNIQUE INDEX UQ_TRG_TRG_EMP_DTTM_CODE ON TRG_TRG
> (TRG_EMP_ID,TRG_DTTM,TRG_CODE);
>
>
> (I shall provide ddl for all dependent tables if required)
>
>
>
> This is more or less by design. The second update in the same
> transaction replaces a delta version (create by the first update)
> with a
> full-length version.
>
> v2.5 is likely to handle it better, but I wouldn't expect any
> miracles...
>
>
> Dmitry
>
>
> Thanks,
> Jaison.
>

[Non-text portions of this message have been removed]