container must be an array or object

[MuhammadTahaNaveed]

Describe the bug
Errors out saying container must be an array or object.

How are you accessing AGE (Command line, driver, etc.)?

• terminal

What data setup do we need to do?

SELECT * FROM cypher('hello',
    $$CREATE (n)-[:KNOWS {n:'hello'}]->({n:'hello'})
$$) as (a agtype);

What is the command that caused the error?

SELECT * FROM cypher('hello',
    $$MATCH (x)<-[y *]-(),({n:y[0].n}) RETURN x
$$) as (a agtype);

ERROR:  container must be an array or object

Expected behavior
expected vertex x to be returned

{"id": 281474976710658, "label": "", "properties": {"n": "hello"}}::vertex

Environment (please complete the following information):

• latest (PG13 master)

Additional Context
This works

SELECT * FROM cypher('hello',$$
    MATCH (x)<-[y *]-()
    MATCH ({n:y[0].n})
    RETURN x
$$) as (a agtype);
                                     a                                      
----------------------------------------------------------------------------
 {"id": 281474976710658, "label": "", "properties": {"n": "hello"}}::vertex
(1 row)

[jrgemignani]

@MuhammadTahaNaveed Looks like I overlooked binary arrays and objects. I will look into this.

Correction: It is because the VLE passes a special vpc container (vpc = VLE Path Container) around for its results. In this case, the container is getting passed instead of the actualized contents of that container. I will see about putting the decode into agtype_access_operator. However, this may not be the appropriate place for it.

[jrgemignani]

@MuhammadTahaNaveed I have fixed it and verified the results with Neo4j. Working on the PR now.

[MuhammadTahaNaveed]

@MuhammadTahaNaveed Looks like I overlooked binary arrays and objects. I will look into this.

Correction: It is because the VLE passes a special vpc container (vpc = VLE Path Container) around for its results. In this case, the container is getting passed instead of the actualized contents of that container. I will see about putting the decode into agtype_access_operator. However, this may not be the appropriate place for it.

@jrgemignani I think the reason you mentioned may also be valid for #1044.

[jrgemignani]

See PR #1046 which addresses this issue.