Welding Natural Language Queries to Analytics IRs with LLMs

Summary: Use LLMs to translate NL analytics queries directly to a custom IR (Weld), making IRs accessible and in some cases improving translation accuracy and end-to-end performance versus SQL-centric nl2sql. nl2weld (GPT-4) uses self-reflection, domain instructions and compiler feedback, attains 77.4% on a Spider subset and emits code 1.2–3x faster than gold SQL/DIN-SQL. (summarized by gpt-5-mini on Feb 09 2026)

Paper ID: 531
Venue: CIDR
Year: 2024
Pagerank: 5.093636e-05
Overall Rank: 11,120 | 23.71%
DOI: -

Incoming Non-self Citations Over Time

No non-self incoming citations found for this paper in this database.

Authors

1. Kaushik Rajan (Microsoft)
2. Aseem Rastogi (Microsoft)
3. Akash Lal (Microsoft)
4. Sampath Rajendra (Microsoft)
5. Krithika Subramanian (Microsoft)
6. Krut Patel (University of Illinois Urbana-Champaign)

BibTeX Citation

@inproceedings{rajan_cidr24,
        address = {Amsterdam, Netherlands},
        series = {{CIDR} '24},
        title = {{Welding Natural Language Queries to Analytics IRs with LLMs}},
        booktitle = {Proceedings of the {Conference} on {Innovative} {Data} {Systems} {Research}},
        author = {Rajan, Kaushik and Rastogi, Aseem and Lal, Akash and Rajendra, Sampath and Subramanian, Krithika and Patel, Krut},
        year = {2024}
}

Incoming Citations (Sorted by Pagerank)

Showing 0 of 0 citing papers.

Rank	Citing Paper	Year	Venue	Pagerank

Outgoing Citations (Sorted by Pagerank)

Showing 12 of 12 cited papers.

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

Rank	Cited Paper	Year	Venue	Pagerank
23	Efficiently Compiling Efficient Query Plans for Modern Hardware	2011	VLDB	0.00054886415
103	DuckDB: an Embeddable Analytical Database	2019	SIGMOD	0.00034161428
534	Building Efficient Query Engines in a High-Level Language	2014	VLDB	0.00017046514
865	Natural language to SQL: Where are we today?	2020	VLDB	0.00013521464
1,229	Weld: A Common Runtime for High Performance Data Analytics	2017	CIDR	0.00011578425
1,444	CatSQL: Towards Real World Natural Language to SQL Applications	2023	VLDB	0.00010769944
1,469	Voodoo - A Vector Algebra for Portable Database Performance on Modern Hardware	2016	VLDB	0.00010678751
4,215	Designing an Open Framework for Query Optimization and Compilation	2022	VLDB	6.8275676e-05
5,765	Charting the Design Space of Query Execution using VOILA	2021	VLDB	6.0953705e-05
6,194	Incorporating Super-Operators in Big-Data Query Optimizers	2020	VLDB	5.9470844e-05
6,253	Declarative Sub-Operators for Universal Data Processing	2023	VLDB	5.940599e-05
7,461	Bringing Compiling Databases to RISC Architectures	2023	VLDB	5.611858e-05

Semantically Similar Papers

#	Overall Rank	Paper	Year	Venue
1	865	Natural language to SQL: Where are we today?	2020	VLDB
2	4,944	Hybrid Querying Over Relational Databases and Large Language Models	2025	CIDR
3	9,305	The Power of Constraints in Natural Language to SQL Translation	2025	VLDB
4	4,045	Logical and Physical Optimizations for SQL Query Execution over Large Language Models	2025	SIGMOD
5	3,787	Combining Small Language Models and Large Language Models for Zero-Shot NL2SQL	2024	VLDB
6	279	Text-to-SQL Empowered by Large Language Models: A Benchmark Evaluation	2024	VLDB
7	10,510	NL2SQLBench: A Modular Benchmarking Framework for LLM-Enabled NL2SQL Solutions	2026	VLDB
8	2,602	NL2SQL is a solved problem... Not!	2024	CIDR
9	9,870	Natural Language to SQL: State of the Art and Open Problems	2025	VLDB
10	2,852	The Dawn of Natural Language to SQL: Are We Fully Ready?	2024	VLDB