Archives of the TeradataForum

Message Posted: Tue, 28 Feb 2006 @ 11:57:40 GMT

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

Try to synchronize the data types of Data_Source_Cd between the big table and you filter table:

     > Data_Source_Cd CHAR(10) CHARACTER SET LATIN NOT CASESPECIFIC,

vs.

     > DATA_SOURCE_CD VARCHAR(30) CHARACTER SET LATIN NOT CASESPECIFIC,

When you join on this column and especially in the context of COALESCE, which is very sensitive to data types, you might force Teradata into an unnecessary data type conversion.

As you probably can't change the definition of the big table, try to change the "filter" definition to the same CHAR(10). It would at least make sense from the data modeling point of view, and, who knows, it might improve performance.

Regards,

Victor

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