Can you explain why workaround is fast. Why one version is fast and
other slow. I'm talking about this change in where:
"and R.ID_ARHDAJ is not null"
to
"and ifnull(R.ID_ARHDAJ,0) <> 0"
For me analyze is no improvement because data is loaded from RDBMS
and would have to run always after load.
Regards Radovan
On 06.12.2019 14:20, Richard Hipp wrote:
On 12/6/19, Keith Medcalf <kmedc...@dessus.com> wrote:
Perhaps the optimizer could make a "left join" that is not actually an outer
join into a "cross join" and preserve the nesting order ... ?
It could do that, but that would kind of defeat the whole purpose of
the optimization, which was to free up the planner to choose a
different nesting order in cases where the LEFT JOIN didn't really
matter.
I suspect that ANALYZE might also solve the OP's problem.
_______________________________________________
sqlite-users mailing list
sqlite-users@mailinglists.sqlite.org
http://mailinglists.sqlite.org/cgi-bin/mailman/listinfo/sqlite-users
