Lightweight Database Wrapper for Unstructured Data.
Tulay Muezzinoglu, Antonio Badia · 2008
We propose an approach to interpreting a structured query language over unstructured data. We define partial records on heterogeneous relations as a means to bridge the gap between structured and unstructured data, as well as an algebra that merges relational operations with Information Extraction operations. This algebra provides support to our Lightweight SQL (L-SQL) language. As part of this effort we define an interface between a relational engine and an Information Extraction module. We have implemented a system based on these ideas and describe the system, including optimization efforts and early experiments. 1.