This repository includes the code base used in the paper "GenJoin: Conditional Generative Plan-to-Plan Query Optimizer that Learns from Subplan Hints", accepted at SIGMOD 2026, the International Conference on Management of Data and available as a preprint on arXiv.
- Pavel Sulimov; Zurich University of Applied Sciences; Winterthur, Switzerland; pavel.sulimov@zhaw.ch
- Claude Lehmann; Zurich University of Applied Sciences; Winterthur, Switzerland; claude.lehmann@zhaw.ch
- Kurt Stockinger; Zurich University of Applied Sciences; Winterthur, Switzerland; kurt.stockinger@zhaw.ch
The repository is split into two separate repositories:
- GenJoin_Repository: This repository holds the code for the GenJoin method, its data collection and scripts for running the experiments. Please read the GenJoin_Repository/README.md for additional information.
- OtherMethods_Repository: This repository contains the code for running AutoSteer and HybridQO. It is based on our previous evaluation benchmark. Please read the OtherMethods_Repository/README.md for additional information.
NOTE: We recommend treating these two folders as distinct repositories, meaning the installed python virtual environment should not be reused across the two directories.
Since we include parts of code bases from recent publications, please make sure to also include their citations. We thank the authors of the previous work for making their research available:
Anneser, Christoph, et al. "Autosteer: Learned query optimization for any sql database." Proceedings of the VLDB Endowment 16.12 (2023): 3515-3527.
Yu, Xiang, et al. "Cost-based or learning-based? A hybrid query optimizer for query plan selection." Proceedings of the VLDB Endowment 15.13 (2022): 3924-3936.
Additionally, we use the Join Order Benchmark published by Leis et al.:
Leis, Viktor, et al. "How good are query optimizers, really?." Proceedings of the VLDB Endowment 9.3 (2015): 204-215.
And the STACK Benchmark published by Markus et al.:
Marcus, Ryan, et al. "Bao: Making learned query optimization practical." Proceedings of the 2021 International Conference on Management of Data. 2021.