# The Case for Learned In-Memory Joins

arXiv: https://arxiv.org/abs/2111.08824 · HTML: https://mlsanity.com/paper/2111.08824/

## Key facts

- Cited by 4 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 16 Nov 2021.

## Built on by (most cited first, 4 of 4)

- [DobLIX: A Dual-Objective Learned Index for Log-Structured Merge Trees](https://mlsanity.com/paper/2502.05369/) (2025)
- [Implementing Semantic Join Operators Efficiently](https://mlsanity.com/paper/2510.08489/) (2025)
- [Baihe: SysML Framework for AI-driven Databases](https://mlsanity.com/paper/2112.14460/) (2021)
- [Sig2Model: A Boosting-Driven Model for Updatable Learned Indexes](https://mlsanity.com/paper/2509.20781/) (2025)

## Abstract (arXiv)

In-memory join is an essential operator in any database engine. It has been extensively investigated in the database literature. In this paper, we study whether exploiting the CDF-based learned models to boost the join performance is practical or not. To the best of our knowledge, we are the first to fill this gap. We investigate the usage of CDF-based partitioning and learned indexes (e.g., Recursive Model Indexes (RMI) and RadixSpline) in the three join categories; indexed nested loop join (INLJ), sort-based joins (SJ) and hash-based joins (HJ). Our study shows that there is a room to improve the performance of INLJ and SJ categories through our proposed optimized learned variants. Our experimental analysis showed that these proposed learned variants of INLJ and SJ consistently outperform the state-of-the-art techniques.