By Franck Pachot . 11gR2 started to have the optimizer react at execution time when a misestimate is encountered. Then the next executions are re-optimized with more accurate estimation, derived from the execution statistics. This was called cardinality feedback. Unfortunately,(…)