6. The rest is usual way to setup datamapper on Merb.
My current implementation is a lot slower than MySQL and sqlite, but changing some data storage strategies speed up the performance dramatically, so there are lots of potential for optimization.
Further research topics/TODO
- Implement outstanding tasks such as multi conditions, other finder conditions (<, <=, like), Data Types
- At this moment, it's only uses B tree, as it covers the wide range of functionality (range search, duplicate values, transaction support, and so on). Consider replacing part to HDB or FDB for compact storage and speed.
- Are there any ways to retrieve first/last key on FDB/HDB?