RE: derived key indexes. end of the CQE?

Not quite, but nearly there!
IBM is doing a great job knocking off SQE restrictions.
There are only a couple of I can think of that remain. One is querying LFs
directly, which appears IBM will remove eventually as well.
The other one is read triggers, but so few people use them, this is only a
minor obstacle to fully going SQE.

This is a very nice article and excellent technology. Hopefully DBAs take
Tom's suggestion and start replacing select/omit LFs with SQL sparse
indexes.

Elvis

Celebrating 11-Years of SQL Performance Excellence on IBM i5/OS and OS/400
www.centerfieldtechnology.com


-----Original Message-----
Subject: derived key indexes. end of the CQE?

a good article on the new derived key indexes in v6r1 in IBM database
magazine:
http://www.ibmdatabasemag.com/story/showArticle.jhtml?articleID=206800753

esp interesting reading at the end of the article. There is a new RCDFMT
clause in the CREATE INDEX statement which allows the created index to be
seen as a logical file for record level i/o.
CREATE INDEX STARTEST.BIGREVENUE ON STARTEST.ITEM_FACT (SUPPKEY ASC,
QUANTITY * EXTENDEDPRICE ASC)
WHERE QUANTITY * EXTENDEDPRICE > 100000
RCDFMT BSUPPREV ADD QUANTITY, EXTENDEDPRICE, REVENUE_WO_TAX, REVENUE_W_TAX;

"...Before V6R1, the existence of DDS-created Select/Omit logical files
would force queries against the underlying table or associated physical file
to be routed to the old Classic Query Optimizer (CQE). You can override that
behavior using the IGNORE_DERIVED_INDEX *YES QAQQINI file option. With V6R1,
queries using the underlying table or physical file from those DDS
Select/Omit logical files, will be routed to the new SQE unless you have
explicitly specified IGNORE_DERIVED_INDEX *NO in your QAQQINI file. The SQE
optimizer will not use DDS Select/Omit logical file indexes. ..."
-Steve