The Key to Effective UDF Optimization: Before Inlining, First Perform Outlining

Summary: Argues full UDF inlining can harm optimizer effectiveness and introduces UDF outlining: decompose and deliberately hide unhelpful UDF fragments before selective inlining to produce optimizer-friendly SQL. PRISM implements this, yielding major gains (avg 1.29× DuckDB, 298.73× SQL Server) via improved unnesting, data skipping, and fewer joins. (summarized by gpt-5-mini on Feb 09 2026)

Paper ID: 13772
Venue: VLDB
Year: 2025
Pagerank: 4.3546206e-05
Overall Rank: 9,343 | 35.01%
DOI: 10.14778/3696435.3696436

Incoming Non-self Citations Over Time

Authors

Incoming Citations (Sorted by Pagerank)

Showing 2 of 2 citing papers.

Rank	Citing Paper	Year	Venue	Pagerank
9,763	The UDFBench Benchmark for General-purpose UDF Queries	2025	VLDB	4.2856106e-05
10,253	Scalable GPU Acceleration of Scalar Functions in Analytical Databases: Compilation, Benchmarking, and Optimization	2026	VLDB	4.1945683e-05

Outgoing Citations (Sorted by Pagerank)

Showing 22 of 22 cited papers.

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

Rank	Cited Paper	Year	Venue	Pagerank
185	DuckDB: an Embeddable Analytical Database	2019	SIGMOD	0.00036538405
418	Morsel-Driven Parallelism: A NUMA-Aware Query Evaluation Framework for the Many-Core Age	2014	SIGMOD	0.00023729211
639	Orthogonal Optimization of Subqueries and Aggregation	2001	SIGMOD	0.00018791492
1,108	Froid: Optimization of Imperative Programs in a Relational Database	2018	VLDB	0.00013984276
1,456	Rewriting Procedures for Batched Bindings	2008	VLDB	0.00011899772
1,582	Execution Strategies for SQL Subqueries	2007	SIGMOD	0.00011265079
1,882	Tuplex: Data Science in Python at Native Code Speed	2021	SIGMOD	0.0001021625
2,237	Procedural Extensions of SQL: Understanding their usage in the wild	2021	VLDB	9.2212748e-05
2,258	SQL Server Column Store Indexes	2011	SIGMOD	9.1678883e-05
2,954	Magpie: Python at Speed and Scale using Cloud Backends	2021	CIDR	7.8262582e-05
3,296	Extracting Equivalent SQL from Imperative Code in Database Applications	2016	SIGMOD	7.2596583e-05
3,648	One WITH RECURSIVE is Worth Many GOTOs	2021	SIGMOD	6.8831123e-05
4,648	Aggify: Lifting the Curse of Cursor Loops using Custom Aggregates	2020	SIGMOD	6.0247446e-05
4,948	Designing an Open Framework for Query Optimization and Compilation	2022	VLDB	5.8116879e-05
5,731	Babelfish: Efficient Execution of Polyglot Queries	2022	VLDB	5.3502065e-05
6,189	Accelerating Python UDFs in Vectorized Query Execution	2022	CIDR	5.1647573e-05
6,375	Dear User-Defined Functions, Inlining isn't working out so great for us. Let's try batching to make our relationship work. Sincerely, SQL	2024	CIDR	5.0923872e-05
6,701	YeSQL: “You extend SQL” with Rich and Highly Performant User-Defined Functions in Relational Databases	2022	VLDB	4.9561066e-05
6,863	Declarative Sub-Operators for Universal Data Processing	2023	VLDB	4.905092e-05
7,448	DBridge: Translating Imperative Code to SQL	2017	SIGMOD	4.7273104e-05
9,718	YeSQL: Rich User-Defined Functions without the Overhead	2022	VLDB	4.2980763e-05
9,719	Tuplex: Robust, Efficient Analytics When Python Rules	2019	VLDB	4.2980763e-05

Semantically Similar Papers

Overall Rank	Paper	Year	Venue	Pagerank
12,316	Fast and Dynamic OLAP Exploration Using UDFs	2009	SIGMOD	4.1945683e-05
4,014	Exploiting Correlations for Expensive Predicate Evaluation	2015	SIGMOD	6.5273084e-05
5,476	Containerized Execution of UDFs: An Experimental Evaluation	2022	VLDB	5.4866534e-05
6,189	Accelerating Python UDFs in Vectorized Query Execution	2022	CIDR	5.1647573e-05
6,645	Functional-Style SQL UDFs With a Capital 'F'	2020	SIGMOD	4.978205e-05
10,459	UDFBench: A Tool for Benchmarking UDF Queries on SQL Engines	2025	SIGMOD	4.1945683e-05
9,763	The UDFBench Benchmark for General-purpose UDF Queries	2025	VLDB	4.2856106e-05
8,583	Efficient Execution of User-Defined Functions in SQL Queries	2023	VLDB	4.4919445e-05
4,582	BlackMagic: Automatic Inlining of Scalar UDFs into SQL Queries with Froid	2019	VLDB	6.070187e-05
6,375	Dear User-Defined Functions, Inlining isn't working out so great for us. Let's try batching to make our relationship work. Sincerely, SQL	2024	CIDR	5.0923872e-05