Guys -
I'm doing a database consistency check for a client and I find that
they're building unique indexes for performance/query reasons where
they could be using non-unique indexes.
Note that these columns in the unique indexes are truly unique and
don't constitute a collision hazard of any kind.
Now, I personally wouldn't use unique where non-unique would do but I
need to know what the performance impact here is.
What's it costing to use unique when non-unique would do?
DB2 8.1.8 on AIX 5.1
Bruce