Dmitry Yemanov
Dmitry Yemanov
> If it was intended as an example, I would assume it is not a robust implementation. It was initially developed by a customer request and was used in production...
What's wrong with the built-in function MOD(A, B)?
It should be already fixed by commit c432bd0, please test the snapshot build.
Likely duplicate for #8136.
The optimizer cannot detect that value 42 matches the index range, as expressions are not evaluated at compile time. You need to specify that explicitly: `SELECT * FROM AUDIT WHERE...
Starting with v5, IN is not decomposed into ORs but instead is processed natively. That said, partial index matching could be improved to handle your case (it was already noted...
It's implemented differently (index keys are not based on expression), but from the optimizer POV they're similar - query expression must be matched to the index expression. The difference is...
In the original message, I suppose reads from `changelogdet` can be reduced using hash anti-join for `NOT EXISTS`, although this is not going to optimize reads from `changelog`. It would...
> 3. Indexes on long varchars can't be created. For example, with single byte charset it's possible to create index on varchar(1024), but not with UTF8 charset. Of course, we...
> I think this is going to cause problems. > > When external routines are created, engine immediately calls external code, that can interact with existing metadata. Hmmm. Maybe theoretically...