it's two bugs
https://bitbucket.org/zzzeek/sqlalchemy/issues/4011/joined-subclass-to-2-level-subquery-load
and reopened in master
https://bitbucket.org/zzzeek/sqlalchemy/issues/3963/aliased-vs-non-aliased-specificity-in
thanks!
On 06/15/2017 01:23 AM, Sherwin Yu wrote:
We found a bug invo
We found a bug involving filtering a JTI relation and subqueryloading. The
emitted SQL for the subqueryload does not correctly join the child and
parent JTI tables, resulting in a cross product.
Info:
- psql (PostgreSQL) 9.5.7
- SQLAlchemy 0.9.10
Suppose we have the following relations