On Mar 20, 2007, at 3:24 PM, Ron Johnson wrote:
> How many *distinct* values are there in int_c? What percentage of
> them match "int_c <= 10"?
right now there are 14, and 80% match -- but next week that number
will be 20 , 60% ... and so on
> Same questions, but for int_h.
same answer
> If int_h is relatively unique, then this index might help:
> (INT_H, INT_C)
If that helps, try
> (INT_H, INT_C, BOOL_A, BOOL_B, BOOL_D, BOOL_E, BOOL_G, BOOL_I)
i'll try both, thanks.
the current 'super index' works - i'm more worried for speed right
now than disk space , so its worth physical overhead.