SPARQL query on OWL data returns empty array

Question

I have this ontology

<owl:NamedIndividual rdf:about="&object_test;TablePick">
    <rdf:type rdf:resource="&dul;PhysicalArtifact"/>
    <SOMA:hasLocalization>
        <owl:NamedIndividual>
            <rdf:type rdf:resource="&SOMA;Localization"/>
            <SOMA:hasSpaceRegion>
                <xyz:pose rdf:resource="&object_test;Pose_TablePick"/>
            </SOMA:hasSpaceRegion>
        </owl:NamedIndividual>
    </SOMA:hasLocalization>
</owl:NamedIndividual>

<owl:NamedIndividual rdf:about="&object_test;Pose_TablePick">
    <rdf:type rdf:resource="&xyz;Pose"/>
    <xyz:quaternion rdf:datatype="&xsd;string">0.0 0.0 1.0 0.0</xyz:quaternion>
    <xyz:translation rdf:datatype="&xsd;string">0.5 0.0 0.2</xyz:translation>
</owl:NamedIndividual>

I am trying to extract the translation and quaternion. This is my initial query is:

SELECT ?obj ?translation WHERE {
  ?obj rdf:type dul:PhysicalArtifact .
  ?pose rdf:type xyz:Pose .
  ?pose xyz:translation ?translation .
}

and I get translation

[object_test.TablePick, '0.5 0.0 0.2']

But I want to multiple queries on "obj", so I changed my query to

SELECT ?obj ?translation WHERE {
  ?obj rdf:type dul:PhysicalArtifact .
}

Since I am using owlready2 I tried to access obj this way

with onto:
  print(obj.hasLocalization[0].hasSpaceRegion[0].translation)

but now the translation is an empty list. How can I access translation in python?

Edit 1: actually the initial query is "wrong" as well, since it gives all possible combinations of objects and locations, not just that particular object's location. So something like this

[[object_test.TablePick, '0.0 0.0 0.0'], [object_test.TablePick, '0.5 0.0 0.2'], [object_test.TablePick, '0.5 -0.1 0.5'], [object_test.TablePick, '0.5 0.1 0.5'], [object_test.obj_1234, '0.0 0.0 0.0'], [object_test.obj_1234, '0.5 0.0 0.2'], [object_test.obj_1234, '0.5 -0.1 0.5'], [object_test.obj_1234, '0.5 0.1 0.5'], [object_test.obj_5678, '0.0 0.0 0.0'], [object_test.obj_5678, '0.5 0.0 0.2'], [object_test.obj_5678, '0.5 -0.1 0.5'], [object_test.obj_5678, '0.5 0.1 0.5']]

Edit 2: "fixed" the sparql query by simplifying the ontology

<SOMA:hasLocalization>
    <owl:NamedIndividual>
        <rdf:type rdf:resource="&SOMA;Localization"/>
        <xyz:pose rdf:resource="&object_test;Pose_obj_5678"/>
    </owl:NamedIndividual>
</SOMA:hasLocalization>

Now I get the correct solution with

[[objects.TablePick, '0.5 0.0 0.2'], [objects.obj_1234, '0.5 -0.1 0.5'], [objects.obj_5678, '0.5 0.1 0.5']]

but trying to access translation with obj in python still gives an empty list.

Edit 3:

print(obj.hasLocalization[0].get_properties())

the result of this is empty. How is that possible when this sparql query works

SELECT ?obj ?translation WHERE {
    ?obj rdf:type dul:PhysicalArtifact .
    ?obj SOMA:hasLocalization ?loc .
    ?loc xyz:pose ?pose .
    ?pose rdf:type xyz:Pose .
    ?pose xyz:translation ?translation .
}