I don't see a lot of chatter about DATE_CORRELATION_OPTIMIZATION (DCO).
Yet, it seems like a really good idea.
Anybody have anything interesting to say about it?
I do have a specific question and a more general question about it.
First, BOL suggests it does best when the dates are the leading or only field in the CK. Well, sure, but that raises the question about how well it does when the date isnot the leading or only field in the CK, especially if there is an FK set up between tables anyway - as is apparently required bfore the DCO starts working anyway.
More generally, my seat of the pants analysis suggests the optimizer could still be usefully alerted to the temporal semantics of two tables and/or queries by the feature anyway - or is that wrong, does the FK do all the work there already? Most generally, I have a situation here where we are setting up a data mart or maybe it's a warehouse where the data all has temporal dimensions, and it has just come home to me what this is going to mean in terms of writing the appropriate SQL and the runtime it will take to execute that SQL.
Thanks.
Josh