The current implementation will select all the fields from the views and SPARQL queries that are joined together and then limit the actual fields that are returned based on the fields specified in the data sources. This may be a performance issue if there are a lot of fields in the views but the service only needs a few of them.
Also, aggregates can be done against fields that are not actually listed in the service metadata which is not the correct behavior.
The current implementation will select all the fields from the views and SPARQL queries that are joined together and then limit the actual fields that are returned based on the fields specified in the data sources. This may be a performance issue if there are a lot of fields in the views but the service only needs a few of them.
Also, aggregates can be done against fields that are not actually listed in the service metadata which is not the correct behavior.