#48 Intelligent Relation Mapping

[Anonymous]

Currently a lot of relations are not represented optimally in the relational database. Based on a statistical model using the database performance, specific database implementation details and sql details, a system could be developed that transparently:
- Assembles statistical information about queries, tables and relations
- Use these statistical data to automagically optimize the relational model

Examples could be:
- Detect 1-many relations based on lists. With this, an additional attribute could be added to the dependent type to relate to the container, if it improves sql performance.
- Use statistical data for index creation (see intelligent indexes)
- Export statistical data through JMX
- Enable hacks for common data query use-cases like:
- Tree usage of objects (get the tree in one query)
- Related objects which seem to be queried together and can be joined
etc