AQORA: A Learned Adaptive Query Optimizer for Spark SQL

2025-10-12

AI Summary

Current query optimization faces two key bottlenecks: learning-based query optimization (LQO) relies on static plans and suffers from low learning efficiency due to the absence of runtime cardinality feedback; adaptive query processing (AQP) over-relies on rule-based heuristics and lacks empirical learning capability. This paper proposes a reinforcement learning–driven adaptive optimization framework for Spark SQL, innovatively unifying LQO and AQP. Our approach incorporates real-execution feature encoding, phased feedback intervention, policy auto-adaptation, and a low-overhead integration mechanism to enable fine-grained, runtime dynamic plan tuning. Experimental results demonstrate that our method reduces end-to-end execution time by up to 90% compared to state-of-the-art learning-based optimizers, and by up to 70% relative to Spark’s default AQP configuration. It significantly improves both optimization efficiency and generalization across diverse workloads.

Technology Category

• Search and Optimization: Learning to Search
• Natural Language Processing: Learning & Optimization for NLP
• Reasoning under Uncertainty: Stochastic Optimization

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Querying, indexing, and retrieval in Web-related graphs
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems

Abstract

Recent studies have identified two main approaches to improve query optimization: learned query optimization (LQO), which generates or selects better query plans before execution based on models trained in advance, and adaptive query processing (AQP), which adapts the query plan during execution based on statistical feedback collected at runtime. Although both approaches have shown promise, they also face critical limitations. LQO must commit to a fixed plan without access to actual cardinalities and typically rely on a single end-to-end feedback signal, making learning inefficient. On the other hand, AQP depends heavily on rule-based heuristics and lacks the ability to learn from experience. In this paper, we present AQORA, an adaptive query optimizer with a reinforcement learning architecture that combines the strengths of both LQO and AQP. AQORA addresses the above challenges through four core strategies: (1) realistic feature encoding, (2) query stage-level feedback and intervention, (3) automatic strategy adaptation, and (4) low-cost integration. Experiments show that AQORA reduces end-to-end execution time by up to 90% compared to other learned methods and by up to 70% compared to Spark SQL's default configuration with adaptive query execution.

Problem

• Combining learned and adaptive query optimization approaches
• Addressing limitations of static plans and runtime adaptation
• Improving Spark SQL performance through reinforcement learning

Innovation

• Reinforcement learning architecture combining LQO and AQP
• Query stage-level feedback with automatic strategy adaptation
• Realistic feature encoding with low-cost integration