for example, screenshot from dbt metrics:
is the same approach efficient for GoodData? Keep database normalized, define metrics, dimensions on gooddata side.
• some databases support join elimination
• databases and GoodData have cache and running the same query many times should be fast
• GoodData has preaggregation feature (beta), probable idea is to keep results of complex calculations in separate tables with aggregated data and instead of running queries - get aggregated data
having all these layers which help to optimize performance - do you prefer star schemas?