"Ross Klatte" <kl********@aol.commmm> wrote in message
news:20***************************@mb-m21.aol.com...
If I have a key comprised of Column-1 plus Column-2, and if I
Select X where Column-2 = "yyy"
then I suppose that the retrieval takes place on a pure
sequential basis just as if I were selecting on some column
not part of any key. Is that right?
Ross
Vontay, Virginia
In the case you described above, DB2 will "usually" do a index scan rather
than a tablespace scan of the data pages. So that means DB2 will read all
the index leaf pages sequentially instead of using the b-tree of the index.
Instead of creating another index on (column-2, column-1) just create a new
index on column-2. Column-1 in this new index would be redundant if you keep
the first index (column-1, column-2).