Re: merge join cartesian- access path?

Tlcbus wrote:
>
> To all: In order to improve performance, I created a small temp table from
> original
> big table to reduce joining operations. In step 4.2 I found merge join
> cartesian step,
> and my query never finishes. generally, cartesian join is caused by missing
> joining
> columns from 2 tables(This is not the case here). Any clues ?
>
> thanks in advance Ted chyn
> ++++++++++++++++++++++++++++++++++++++++++++++
> plan after creation of small temp work table
>
> 1.5331 SELECT STATEMENT Cost = 5331
> 2.1 SORT GROUP BY
> 3.1 HASH JOIN
> 4.1 TABLE ACCESS FULL TEMP_WOSN_HEADER
> 4.2 MERGE JOIN CARTESIAN =====What is this means ?
> 5.1 VIEW
> 6.1 SORT AGGREGATE
> 7.1 TABLE ACCESS BY ROWID T_UTL_DATES
> 8.1 INDEX UNIQUE SCAN I_PK_UDAT_CALENDAR_DY UNIQUE
> 5.2 TABLE ACCESS BY ROWID T_WOSN_OPEN
> 6.1 INDEX RANGE SCAN I_WSNO_UCD_DY NON-UNIQUE
> ==================================================
> original query - use nest loop join
> 1. SELECT STATEMENT Cost =
> 2.1 SORT GROUP BY
> 3.1 NESTED LOOPS
> 4.1 NESTED LOOPS
> 5.1 VIEW
> 6.1 SORT AGGREGATE
> 7.1 TABLE ACCESS BY ROWID T_UTL_DATES
> 8.1 INDEX UNIQUE SCAN I_PK_UDAT_CALENDAR_DY UNIQUE
> 5.2 TABLE ACCESS BY ROWID T_WOSN_OPEN
> 6.1 INDEX RANGE SCAN I_WSNO_UCD_DY NON-UNIQUE
> 4.2 TABLE ACCESS BY ROWID T_WOSN_HEADER
> 5.1 INDEX RANGE SCAN I_WSNH_WOSN_CD NON-UNIQUE
>

Hi,

without publishing your SQL-Statement no one can help you.

-- 
Regards

M.Gresz    :-)