LINDAS: a learned approach to index algorithm selection

Chaohong Ma, Xiaohui Yu, Yifan Li, Aishan Maoliniyazi, Xiaofeng Meng

Abstract: The recent surge in learned index algorithms, alongside traditional indexes, has created a diverse and extensive array of indexing options to support query processing in database systems. Despite the rapid expansion of learned indexes, there remains a significant gap in tools for index algorithm selection. Traditional research on index selection has largely focused on recommending which columns to index, as the choice between algorithms like B+tree or hash index was once straightforward. This was managed through basic rules or experiential judgment, given the historically limited options. However, this approach is inadequate today, due to the growing diversity and complexity of index algorithms. In this paper, we introduce a Learned INDex Algorithm Selector, LINDAS. Taking a learned approach, LINDAS uniquely focuses on automatically selecting the most suitable index algorithm (e.g., a traditional index or one of the recently proposed learned indexes) for a specific column that satisfies diverse performance objectives in a wide range of application scenarios. We explore the design space of LINDAS, employing a carefully designed featurization approach to capture both data- and workload-specific characteristics with attention mechanisms, as well as the meta-features of index algorithms. Two variants (clf and reg) of LINDAS are designed to cater to diverse applications and adapt readily to new datasets, workloads, and emerging index algorithms. We conduct comprehensive evaluations of LINDAS across various datasets and workloads, demonstrating its effectiveness and superiority compared to applicable baselines. We highlight interesting problems that merit future investigation in index algorithm selection, and propose potential solutions.

External IDs:dblp:journals/kais/MaYLMM25