Re: [Sqlalchemy-tickets] [sqlalchemy] #2587: flatten joined-inheritance join targets on individual tables

[sqlalchemy <mi...@zz...>]

#2587: flatten joined-inheritance join targets on individual tables
------------------------------+---------------------------------------
      Reporter:  zzzeek       |      Owner:  zzzeek
          Type:  enhancement  |     Status:  new
      Priority:  medium       |  Milestone:  0.9.0
     Component:  orm          |   Severity:  very major - up to 2 days
    Resolution:               |   Keywords:
Progress State:  in queue     |
------------------------------+---------------------------------------

Comment (by zzzeek):

 also:

 when we join from A to a mapper "B join C" and we are aliasing, as in
 eager loading:

 {{{
 orm_join(A, BC.alias())

 SELECT a.id AS a_id, anon_1.b_id AS anon_1_b_id, anon_1.c_id AS
 anon_1_c_id
 FROM a JOIN (
    SELECT b.id AS b_id, c.id AS c_id FROM
    b JOIN c ON b.id=c.id
 ) AS anon_1 ON a.id=anon_1.b_id

 }}}

 we can do the "flattened" thing, where we don't create the subquery, but
 again produce the same labels:

 {{{
 orm_join(A, BC.unnessted_alias())

 SELECT a.id AS a_id, anon_1_b.id AS anon_1_b_id, anon_1_c.id AS
 anon_1_c_id
 FROM a JOIN (b AS anon_1_b JOIN c AS anon_1_c ON anon_1_b.id=anon_1_c.id)
 ON a.id=anon_1_b_id

 }}}

 again, we get the same labeling conventions and everything, we here just
 change how `Alias` represents its contents.  Perhaps `JoinedAlias`
 subclass.

 The above query, when run through the previous compiler "subquery the
 nested joins" converter, gets:

 {{{

 SELECT a.id AS a_id, anon_2.anon_1_b_id AS anon_1_b_id, anon_2.anon_1_c_id
 AS anon_1_c_id
 FROM a JOIN (
      SELECT anon_1_b.id AS anon_1_b_id, anon_1_c.id AS anon_1_c_id
      FROM b AS anon_1_b JOIN c AS anon_1_c ON anon_1_b.id=anon_1_c.id
 ) AS anon_2
 ON a.id=anon_2.anon_1_b_id


 }}}

 and it works

--
Ticket URL: <http://www.sqlalchemy.org/trac/ticket/2587#comment:4>
sqlalchemy <http://www.sqlalchemy.org/>
The Database Toolkit for Python