QTrail-DB: A Query Processing Engine for Imperfect Databases with Evolving Qualities
share
Imperfect databases are very common in many applications due to various reasons ranging from data-entry errors, transmission or integration errors, and wrong instruments' readings, to faulty experimental setups leading to incorrect results. The management and query processing of imperfect databases is a very challenging problem as it requires incorporating the data's qualities within the database engine. Even more challenging, the qualities are typically not static and may evolve over time. Unfortunately, most of the state-of-art techniques deal with the data quality problem as an offline task that is in total isolation of the query processing engine (carried out outside the DBMS). Hence, end-users will receive the queries' results with no clue on whether or not the results can be trusted for further analysis and decision making. In this paper, we propose the it "QTrail-DB" system that fundamentally extends the standard DBMSs to support imperfect databases with evolving qualities. QTrail-DB introduces a new quality model based on the new concept of "Quality Trails", which captures the evolution of the data's qualities over time. QTrail-DB extends the relational data model to incorporate the quality trails within the database system. We propose a new query algebra, called "QTrail Algebra", that enables seamless and transparent propagation and derivations of the data's qualities within a query pipeline. As a result, a query's answer will be automatically annotated with quality-related information at the tuple level. QTrail-DB propagation model leverages the thoroughly-studied propagation semantics present in the DB provenance and lineage tracking literature, and thus there is no need for developing a new query optimizer. QTrail-DB is developed within PostgreSQL and experimentally evaluated using real-world datasets to demonstrate its efficiency and practicality.
READ FULL TEXT
