The optimization of select-project-join (SPJ) queries entails two major challenges: (i) finding a good join order and (ii) selecting the best-fitting physical join operator for each single join within the chosen join order. Previous work mainly focuses on the computation of a good join order, but leaves open to which extent the physical join operator selection accounts for plan quality. Our analysis using different query optimizers indicates that physical join operator selection is crucial and that none of the investigated query optimizers reaches the full potential of optimal operator selections. To unlock this potential, we propose TONIC , a novel cardinality estimation-free extension for generic SPJ query optimizers in this paper. TONIC follows a learning-based approach and revises operator decisions for arbitrary join paths based on learned query feedback. To continuously capture and reuse optimal operator selections, we introduce a lightweight yet powerful Query Execution Plan Synopsis ( QEP-S ). In comparison to related work, TONIC enables transparent planning decisions with consistent performance improvements. Using two real-life benchmarks, we demonstrate that extending existing optimizers with TONIC substantially reduces query response times with a cumulative speedup of up to 2.8x.

Every year more and more advanced approaches to cardinality estimation are published, using learned models or other data and workload specific synopses. In contrast, the majority of commercial in-memory systems still relies on sampling. It is arguably the most general and easiest estimator to implement. While most methods do not seem to improve much over sampling-based estimators in the presence of non-selective queries, sampling struggles with highly selective queries due to limitations of the sample size. Especially in situations where no sample tuple qualifies, optimizers fall back to basic heuristics that ignore attribute correlations and lead to large estimation errors. In this work, we present a novel approach, dealing with these 0-Tuple Situations. It is ready to use in any DBMS capable of sampling, showing a negligible impact on optimization time. Our experiments on real world and synthetic data sets demonstrate up to two orders of magnitude reduced estimation errors. Enumerating single filter predicates according to our estimates reveals 1.3 to 1.8 times faster query responses for complex filters.