To Join or Not to Join? Thinking Twice about Joins before Feature Selection

Summary: Safe-join avoidance for feature selection in normalized datasets: many join-derived features can be dropped without hurting ML accuracy. Experiments on real normalized datasets show accurate safety predictions and substantial runtime savings for popular feature selection methods. (summarized by gpt-5-nano on Feb 09 2026)

Paper ID: h543d7bd254547c72
Venue: SIGMOD
Year: 2016
Pagerank: 0.00014048128
Overall Rank: 779 | 94.77%
DOI: 10.1145/2882903.2882952

Incoming Non-self Citations Over Time

Authors

1. Arun Kumar (University of Wisconsin)
2. Jeffrey Naughton (University of Wisconsin)
3. Jignesh M. Patel (University of Wisconsin)
4. Xiaojin Zhu (University of Wisconsin)

BibTeX Citation

@inproceedings{kumar_sigmod16,
        title = {{To Join or Not to Join? Thinking Twice about Joins before Feature Selection}},
        author = {Kumar, Arun and Naughton, Jeffrey and Patel, Jignesh M. and Zhu, Xiaojin},
        series = {{SIGMOD} '16},
        booktitle = {Proceedings of the {ACM} {SIGMOD} International Conference on Management of Data},
        publisher = {Association for Computing Machinery},
        doi = {10.1145/2882903.2882952},
        url = {https://dl.acm.org/doi/10.1145/2882903.2882952},
        year = {2016}
}

Incoming Citations (Sorted by Pagerank)

Showing 41 of 41 citing papers.

Rank	Citing Paper	Year	Venue	Pagerank
1,038	ARDA: Automatic Relational Data Augmentation for Machine Learning	2020	VLDB	0.000123653
1,153	Data Management Challenges in Production Machine Learning	2017	SIGMOD	0.00011793347
1,223	Data Management in Machine Learning: Challenges, Techniques, and Systems	2017	SIGMOD	0.00011468426
1,257	Towards Linear Algebra over Normalized Data	2017	VLDB	0.0001130959
1,293	Finding Related Tables in Data Lakes for Interactive Data Science	2020	SIGMOD	0.00011144991
1,568	HELIX: Holistic Optimization for Accelerating Iterative Machine Learning	2019	VLDB	0.00010208225
2,162	DIFF: A Relational Interface for Large-Scale Data Explanation	2019	VLDB	8.9344773e-05
2,201	Enabling and Optimizing Non-linear Feature Interactions in Factorized Linear Algebra	2019	SIGMOD	8.8708356e-05
2,663	A Layered Aggregate Engine for Analytics Workloads	2019	SIGMOD	8.1542952e-05
2,802	Distributed Join Algorithms on Thousands of Cores	2017	VLDB	7.9866934e-05
2,908	AI Meets Database: AI4DB and DB4AI	2021	SIGMOD	7.8716173e-05
2,978	In-Database Learning with Sparse Tensors	2018	PODS	7.7872011e-05
3,016	Data Acquisition for Improving Machine Learning Models	2021	VLDB	7.7492494e-05
3,521	Ember: No-Code Context Enrichment via Similarity-Based Keyless Joins	2022	VLDB	7.2320843e-05
3,617	Leva: Boosting Machine Learning Performance with Relational Embedding Data Augmentation	2022	SIGMOD	7.1540505e-05
3,673	Are Key-Foreign Key Joins Safe to Avoid when Learning High-Capacity Classifiers?	2018	VLDB	7.1075403e-05
4,411	LakeBench: A Benchmark for Discovering Joinable and Unionable Tables in Data Lakes	2024	VLDB	6.6089506e-05
4,902	Scalable Asynchronous Gradient Descent Optimization for Out-of-Core Models	2017	VLDB	6.3617262e-05
5,078	Rotom: A Meta-Learned Data Augmentation Framework for Entity Matching, Data Cleaning, Text Classification, and Beyond	2021	SIGMOD	6.2832055e-05
5,279	MATE: Multi-Attribute Table Extraction	2022	VLDB	6.1973533e-05
5,533	Putting Things into Context: Rich Explanations for Query Answers using Join Graphs	2021	SIGMOD	6.0906707e-05
5,748	Optimizing Data Acquisition to Enhance Machine Learning Performance	2024	VLDB	6.0054988e-05
5,877	BlinkML: Efficient Maximum Likelihood Estimation with Probabilistic Guarantees	2019	SIGMOD	5.9599042e-05
6,146	Efficient Construction of Approximate Ad-Hoc ML models Through Materialization and Reuse	2018	VLDB	5.869373e-05
6,594	Tuple-oriented Compression for Large-scale Mini-batch Stochastic Gradient Descent	2019	SIGMOD	5.7394502e-05
6,747	Mitigating the Impedance Mismatch between Prediction Query Execution and Database Engine	2025	SIGMOD	5.6883491e-05
6,775	Causal Feature Selection for Algorithmic Fairness	2022	SIGMOD	5.6838053e-05
7,203	Coresets over Multiple Tables for Feature-rich and Data-efficient Machine Learning	2023	VLDB	5.5845207e-05
7,872	Mind the Gap: Bridging Multi-Domain Query Workloads with EmptyHeaded	2017	VLDB	5.4339723e-05
8,501	SPRINTER: A Fast n-ary Join Query Processing Method for Complex OLAP Queries	2020	SIGMOD	5.3285575e-05
8,815	ApproxML: Efficient Approximate Ad-Hoc ML Models Through Materialization and Reuse	2019	VLDB	5.2703166e-05
9,261	Leveraging Similarity Joins for Signal Reconstruction	2018	VLDB	5.2032182e-05
10,457	Eliminating Redundant Feature Tests in Decision Tree and Random Forest Inference on SQL Predicates	2026	SIGMOD	4.9769913e-05
10,463	Factorized and Vectorized Execution: Optimizing Analytical and Semantic Queries over Relations	2026	SIGMOD	4.9769913e-05
10,664	InferF: Declarative Factorization of AI/ML Inferences over Joins	2026	SIGMOD	4.9769913e-05
10,749	Database Views as Explanations for Relational Deep Learning	2026	VLDB	4.9769913e-05
11,584	Relational Query Synthesis ⋈ Decision Tree Learning	2024	VLDB	4.9769913e-05
11,596	Enriching Relations with Additional Attributes for ER	2024	VLDB	4.9769913e-05
11,712	Regularized Pairwise Relationship based Analytics for Structured Data	2023	SIGMOD	4.9769913e-05
11,987	Enforcing Constraints for Machine Learning Systems via Declarative Feature Selection: An Experimental Study	2021	SIGMOD	4.9769913e-05
12,253	Learning Efficiently Over Heterogeneous Databases	2018	VLDB	4.9769913e-05

Outgoing Citations (Sorted by Pagerank)

Showing 8 of 8 cited papers.

Citations counted here include only citations to other VLDB/SIGMOD/CIDR/PODS papers in this database.

Rank	Cited Paper	Year	Venue	Pagerank
105	The MADlib Analytics Library or MAD Skills, the SQL	2012	VLDB	0.00033633007
473	Skew-Aware Automatic Database Partitioning in Shared-Nothing, Parallel OLTP Systems	2012	SIGMOD	0.00017684278
537	MLbase: A Distributed Machine-learning System	2013	CIDR	0.00016762761
654	Materialization Optimizations for Feature Selection Workloads	2014	SIGMOD	0.00015096817
731	Learning Generalized Linear Models Over Normalized Data	2015	SIGMOD	0.00014400356
1,009	Simulation of Database-Valued Markov Chains Using SimSQL	2013	SIGMOD	0.0001253551
2,587	Brainwash: A Data System for Feature Engineering	2013	CIDR	8.2523942e-05
7,542	Feature Selection in Enterprise Analytics: A Demonstration using an R-based Data Analytics System	2013	VLDB	5.4976824e-05

Semantically Similar Papers

#	Overall Rank	Paper	Year	Venue
1	5,609	On Producing Join Results Early	2003	PODS
2	8,402	The Case for Learned In-Memory Joins	2023	VLDB
3	5,774	Optimizing Data Pipelines for Machine Learning in Feature Stores	2023	VLDB
4	2,844	FactorJoin: A New Cardinality Estimation Framework for Join Queries	2023	SIGMOD
5	2,613	Auto-Join: Joining Tables by Leveraging Transformations	2017	VLDB
6	10,199	Featurized-Decomposition Join: Low-Cost Semantic Joins with Guarantees	2026	VLDB
7	7,203	Coresets over Multiple Tables for Feature-rich and Data-efficient Machine Learning	2023	VLDB
8	1,257	Towards Linear Algebra over Normalized Data	2017	VLDB
9	3,673	Are Key-Foreign Key Joins Safe to Avoid when Learning High-Capacity Classifiers?	2018	VLDB
10	731	Learning Generalized Linear Models Over Normalized Data	2015	SIGMOD