At the moment dataset_id is used for 2 things:
Identify the id of a dataset:
dataset -dataset_id->id
Identify the dataset which a variable belongs to:
variable-dataset_id->id
This is ambiguous, because the same property is used to link something to a dataset id, and the meaning is different. In the first case, we are describing that a dataset has an id; in the second one we are saying that a variable belongs to a dataset.
We had agreed we would use P2006020004 (dataset) instead in the second case. But I see it's not been done in the code.
At the moment dataset_id is used for 2 things: Identify the id of a dataset:
dataset -dataset_id->id
Identify the dataset which a variable belongs to:
variable-dataset_id->id
This is ambiguous, because the same property is used to link something to a dataset id, and the meaning is different. In the first case, we are describing that a dataset has an id; in the second one we are saying that a variable belongs to a dataset.
We had agreed we would use P2006020004 (dataset) instead in the second case. But I see it's not been done in the code.