Some drivers do automatic aggregation of selected measures based on the selected dimensions, however this breaks clients that aren't expecting the cardinality of the result to change based on the select list (and are caching results and just querying additional/missing values).
-metrics and dimensions. Select * lists metrics and dimensions; result set implicitly groups by metrics
-sqlcolumns would differentiate between metrics and dimensions
-sqltables would need to differentiate between tables and cubes
-issues around semantics; does filter do pre- or post-filtering?
-alternatively add syntax for doing operations against cubes?
Some drivers do automatic aggregation of selected measures based on the selected dimensions, however this breaks clients that aren't expecting the cardinality of the result to change based on the select list (and are caching results and just querying additional/missing values).
-metrics and dimensions. Select * lists metrics and dimensions; result set implicitly groups by metrics -sqlcolumns would differentiate between metrics and dimensions -sqltables would need to differentiate between tables and cubes -issues around semantics; does filter do pre- or post-filtering? -alternatively add syntax for doing operations against cubes?