Density nonsense

[Nicolas Dumeige]

Eric,

During batch, the FK would be most painful, and again, what could possibly be the gain ?

Referencial constraint serve no purpose here.
The index on a 2 value table neither.

Thanks for your help anyway.

[Jean-Luc Oudart]

In Absolute , not so bad
But compared to the overall size of your table, you wonder about the density "nonsense", I would think that the sample size speaks for itself.

The overall distribution is 93% / 7% for 840M records. But 420,000 would you expect this distribution to be uniform ?

From the overview on histograms :
When to Not Use Histograms
--------------------------

Also, be aware that histograms, as well as all optimizer statistics, are
static. If the data distribution of a column changes frequently, it is
necessary to recompute the histogram for a given column. Histograms are not
useful for columns with the following characteristics:

o all predicates on the column use bind variables
o the column data is uniformly distributed
o the column is not used in WHERE clauses of queries
o the column is unique and is used only with equality predicates


Regards
Jean-Luc

[Nicolas Dumeige]

FYI : all conditions are met.

For what I read in the statistics in 20 GB data sample, Oracle have found 0.51 record with a specific value and that is a nonesense.
And the histograms is studpid as weel with value of a such magnitude 10E35 !

Furthermore, I have 4 others columns with the same kind of crap ...

Thanks all for your post.