# Synthesizing Mapping Relationships Using Table Corpus

arXiv: https://arxiv.org/abs/1705.09276 · HTML: https://mlsanity.com/paper/1705.09276/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 25 May 2017.

## Built on by (most cited first, 2 of 2)

- [MMTU: A Massive Multi-Task Table Understanding and Reasoning Benchmark](https://mlsanity.com/paper/2506.05587/) (2025)
- [TabulaX: Leveraging Large Language Models for Multi-Class Table Transformations](https://mlsanity.com/paper/2411.17110/) (2024)

## Abstract (arXiv)

Mapping relationships, such as (country, country-code) or (company, stock-ticker), are versatile data assets for an array of applications in data cleaning and data integration like auto-correction and auto-join. However, today there are no good repositories of mapping tables that can enable these intelligent applications. Given a corpus of tables such as web tables or spreadsheet tables, we observe that values of these mappings often exist in pairs of columns in same tables. Motivated by their broad applicability, we study the problem of synthesizing mapping relationships using a large table corpus. Our synthesis process leverages compatibility of tables based on co-occurrence statistics, as well as constraints such as functional dependency. Experiment results using web tables and enterprise spreadsheets suggest that the proposed approach can produce high quality mappings.