A Hadoop-Based Database Querying Approach for Non-expert Users

Abstract

Increasingly complex relational databases are challenging for non-expert users to query. To facilitate querying, keyword search and form-based interface methods have been proposed, but they both have limitations. Besides, existing methods rely on single machine, leading to insufficient memory and long time consuming for large-scale data. This paper proposes a Hadoop-based database querying method combining the advantages of form-based interface and keyword search. We first generate a set of query forms according to the database schema, and build distributed indexes offline. At online query stage, aiming at the keywords that users input, we return top-k related forms. After users choose and fill out the form, we can automatically create a structured query and return results to them. Experiments show that our method can not only provide an easy way to query databases for non-expert users, but also achieve better performance in time efficiency.