A query with projection returns doubles

Question

A query with projection returns doubles

Closed this issue 6 months ago · 4 comments

A query with projection returns a list of entities that contains doubles, same entities repeated many times.
Simple example:
result_1 = MyKind.query(ancestor=ancestor_key).fetch()
returns 6 entities, while
result_2 = MyKind.query(ancestor=ancestor_key).fetch(projection=["attr1", "attr2"])
returns 8 entities.

Doing [entity.key.id() for entity in result_2], we find that the query with projection has returned doubles, entities that are repeated in the result set.

This does not seem to be a documented behaviour.

Answer 1 · 2024-02-24T06:22:31.000Z

This is documented behaviour at https://cloud.google.com/appengine/docs/legacy/standard/python/ndb/projectionqueries#Projections_and_multiple_valued_properties .

Answer 2 · 2024-02-24T06:23:02.000Z

Specifically "Projecting a property with multiple values will not populate all values for that property. Instead, a separate entity will be returned for each unique combination of projected values matching the query"

Answer 3 · 2024-02-24T06:25:36.000Z

Also documented at https://cloud.google.com/datastore/docs/concepts/queries#projections_and_array-valued_properties .

Answer 4 · 2024-02-25T13:42:39.000Z

Thanks a lot for the investigation and for the answer. So it is documented indeed. It looks like a weird design choice to me, but at least we know now.