On Fri, 26 Apr 2013 14:06:23 +0200
Clemens Ladisch <[email protected]> wrote:
> When using CROSS where standard SQL allows it (for an actual cartesian
> product), the table ordering does not really matter for optimization
> purposes because the DB has to do two nested full table scans anyway.
Well, no, not really: it doesn't *have* to.
Logically,
from A join B
on A.a = B.a
is exactly the same as
from A cross join B
where A.a = B.a
and
from A, B
where A.a = B.a
and the DBMS is free to treat them the same. In fact, many do. :-)
--jkl
_______________________________________________
sqlite-users mailing list
[email protected]
http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
