Archives of the TeradataForum

Message Posted: Tue, 26 Nov 2002 @ 22:46:17 GMT

	<Prev	Next>		<<First	<Prev	Next>	Last>>

Anomy Anom wrote:

If this is part of an automated process to collect statistics on the primary index, you might adopt a technique that could save you some work. I normally create my tables so that the PI has a standard name (I use 'PIDX'), for example:

     CREATE SET TABLE TEMP.TBL_1
        ( COL_1  INTEGER
        , COL_2  SMALLINT
        , COL_3  DATE
        )

        PRIMARY INDEX PIDX (COL_1, COL_2)
        UNIQUE INDEX SIDX_1 (COL_3);

Then I can use the following COLLECT statements:

     COLLECT STATISTICS ON TEMP.TBL_1 INDEX PIDX;
     COLLECT STATISTICS ON TEMP.TBL_1 INDEX SIDX_1;

If the PI for the table is re-defined, then there's no need to change the COLLECT statement.

	<Prev	Next>		<<First	<Prev	Next>	Last>>