Archives of the TeradataForum

Message Posted: Mon, 16 Oct 2006 @ 21:55:49 GMT

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

If a JI contains rows where some columns have repeating values then you can indicate this in the Create JI DDL by grouping the columns with repeating information into one group & grouping the columns without repeating information into another group something like...

     create join index temp_ji as sel (repeat_col1, repeat_col2,...),
     (other_col_1, other_col_2,...)
     from table1 a inner join table2 b on a.cola = b.colb;

In this case, The created JI would be compressed. For all the rows with the same repeating part, There would be only one physical row having the repeating part stored only once with the remaining part of all the rows from its group.

In this context, That single row is the physical row which can have several logical rows.

Regards,

Sanjaya

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