python - SQLAlchemy memory usage: Can I tell SQLAlchemy to NOT keep track of entities in the session?

Question

I some times use the SQLAlchemy ORM layer for read-only operations, typically for some sort of reporting task, something like:

for item in session.query(Item):
    for relatedobject in item.relatedobjects:
        print(item.id, item.name, relatedobject.id, relatedobject.name)

In this case:

• I never modify the entities in the unit of work, and
• The item -> relatedobject relationship is one-to-many (i.e. a relatedobject always belongs to a single item entity)

Question:

• Unless I am mistaken, all entities read from the db in the case above is tracked in memory by the SQLAlchemy session (for change tracking etc.)
• For some of these reporting loads, the python process uses a lot of memory due to this in-mem tracking (I believe)
• Given the case above, is there a way I can tell SQLAlchemy to not keep track of the entities in memory during the operation?

I do not want to use the SQLAlchemy expression layer, since I want to use the relationship functionality in the ORM layer.

question from:https://stackoverflow.com/questions/65918721/sqlalchemy-memory-usage-can-i-tell-sqlalchemy-to-not-keep-track-of-entities-in

Answer 1

Waitting for answers