In fact I've used explicit resultMaps in my tests
which should add very little overhead
Still performance was noticeable different (around 10%). OK, I'll try to do more profiling these days to be confident what is the difference and where does toUpperString gets called. Thanks
