Query Performance - PI & SI usage


Query Performance - PI & SI usage


  I have two tables that I need to join:

Table1 - CT t1 (col1, col2, col3, col4, col5, col6, col7,col8,col9)

PI - col1, col2,col7,col8

Table2 - CT t2 (col11, col12, col13, col14, col15, col16, col17,col18,col19)

PI - col11, col12

Table2 JOIN Table1

ON t1.col1 = t2.col13

and t1.col2 = t2.col14

and t1.col7 = t2.col18

and t1.col8 = t2.col19

Table2 - big table (history)

Table1 - delta records

Now since the PI of the tables are different would it be helpful if I define an SI on the columns (col13,col14) or all the joining columns? I cannot include col 13, col14 in the PI as it will affect the dirstibution of the table since these values will contain 30% null. Is there any other approach I can consider in order to join these tables in an effecient manner.