On Thu, Oct 12, 2006 at 08:34:45AM +0200, Florian Weimer wrote:
> Some statistics are very hard to gather from a sample, e.g. the number
> of distinct values in a column.

Then how can the DBA know it, either?  The problem with this sort of
argument is always that people are claiming some special knowledge is
available to the DBA.  If it's true that the DBA really _can_ know
this stuff, then there must be some way to learn it.  Which means
that you can, in principle, figure out ways to communicate that to
the optimizer.

I like the suggestion, though, that there be ways to codify known
relationships in the system in such a way that the optimizer can
learn to use that information.  _That_ seems to me to be a big
improvement, because it can be taken into consideration along with
relationships that emerge from the statistics, that the DBA may not
know about.


Andrew Sullivan  | [EMAIL PROTECTED]
In the future this spectacle of the middle classes shocking the avant-
garde will probably become the textbook definition of Postmodernism. 
                --Brad Holland

---------------------------(end of broadcast)---------------------------
TIP 2: Don't 'kill -9' the postmaster

Reply via email to