SQLAlchemy and WITH ... DELETE FROM

SQL:

WITH mt_to_clean AS (
	SELECT mt.key
	FROM
		master_table AS mt
			LEFT JOIN secondary_table_1 AS position
				ON mt.key = position.mt_key
			LEFT JOIN secondary_table_2 AS bvd_copy
				ON mt.key = bvd_copy.mt_key
	WHERE
		position.mt_key IS NULL
		AND
		bvd_copy.mt_key IS NULL
)
DELETE FROM mt
WHERE mt.key = mt_to_clean.key

OR

DELETE FROM mt
WHERE mt.key IN (
	SELECT mt.key
	FROM
		master_table AS mt
			LEFT JOIN secondary_table_1 AS position
				ON mt.key = position.mt_key
			LEFT JOIN secondary_table_2 AS bvd_copy
				ON mt.key = bvd_copy.mt_key
	WHERE
		position.mt_key IS NULL
		AND
		bvd_copy.mt_key IS NULL
)

----------------------
SQLAlchemy:
unreachable_mt_query = session.query(MT.key.label('mt_key')) \
    .outerjoin(SecondaryTable1) \
    .outerjoin(SecondaryTable2) \
    .filter(
        SecondaryTable1.mt_key.is_(None),
        SecondaryTable2.mt_key.is_(None),
    )

unreachable_mt_alias = sa.orm.aliased(unreachable_mt_query.cte(),
                                      name='unreachable_mt')

session.query(MT) \
    .join(unreachable_mt_alias,
          unreachable_mt_alias.c.mt_key == MT.key) \
    .delete(synchronize_session=False)

OR

session.query(MT) \
	.outerjoin(SecondaryTable1) \
	.outerjoin(SecondaryTable2) \
	.filter(
	    SecondaryTable1.mt_key.is_(None),
	    SecondaryTable2.mt_key.is_(None),
	) \
	.delete(synchronize_session=False)

both generate this exception:

Traceback (most recent call last):
  File "/Users/gilles/devel/bla/venv/bin/purge-entries", line 9, in <module>
    load_entry_point('bla==0.6.3', 'console_scripts', 'purge-entries')()
  File "/Users/gilles/devel/bla/bla/tasks.py", line 165, in main
    purge_db(int(args.days))
  File "/Users/gilles/devel/bla/bla/tasks.py", line 68, in purge_db
    for obj in purge_spurious_entries(session):
  File "/Users/gilles/devel/bla/bla/purge_entries.py", line 257, in purge_spurious_entries
    .delete(synchronize_session=False)
  File "/Users/gilles/devel/bla/venv/lib/python2.7/site-packages/sqlalchemy/orm/query.py", line 2827, in delete
    self, synchronize_session)
  File "/Users/gilles/devel/bla/venv/lib/python2.7/site-packages/sqlalchemy/orm/persistence.py", line 1275, in factory
    }, synchronize_session, query)
  File "/Users/gilles/devel/bla/venv/lib/python2.7/site-packages/sqlalchemy/orm/persistence.py", line 1101, in _factory
    return klass(*arg)
  File "/Users/gilles/devel/bla/venv/lib/python2.7/site-packages/sqlalchemy/orm/persistence.py", line 1267, in __init__
    super(BulkDelete, self).__init__(query)
  File "/Users/gilles/devel/bla/venv/lib/python2.7/site-packages/sqlalchemy/orm/persistence.py", line 1066, in __init__
    self._validate_query_state()
  File "/Users/gilles/devel/bla/venv/lib/python2.7/site-packages/sqlalchemy/orm/persistence.py", line 1084, in _validate_query_state
    (methname, )
sqlalchemy.exc.InvalidRequestError: Can't call Query.update() or Query.delete() when join(), outerjoin(), select_from(), or from_self() has been called

if this gives you all results you want to be deleted:
session.query(MT.id).outerjoin(ST1).outerjoin(ST2).filter(ST1.mt_key.is_(None),ST2.mt_key.is_(None)).all()

this is supposed to delete them:
session.query(MT).outerjoin(ST1).outerjoin(ST2).filter(ST1.mt_key.is_(None),ST2.mt_key.is_(None)).delete(synchronize_session=False)
use DB level cascades if need to propagate deletes

apparently in 1.0 it's not supposed to work because of this

try:

q = session.query(MT.id).outerjoin(ST1).outerjoin(ST2).filter(ST1.mt_key.is_(None),ST2.mt_key.is_(None)).subquery()
session.query(MT).filter(MT.id.in_(q)).delete()

Well it did not work previously as it indeed failed to construct the query discarding the bits I was interested in.
The other form I expected to work:

session.query(MT).filter(MT.key.in_(unreachable_mt_query.subquery()))

Did not work with 0.9.8 as well because the table is a polymorphic type and the subquery returned extra columns. This is now fixed in 1.0 it seems. Thank you so much for your help.