Big data have wide applications. Since the quality of big data plays a crucial role in these data-centric applications, data quality management techniques for big data are in demand. Although some theories and techniques for data quality management have been proposed, due to the volume, variety and velocity of big data, current methods could hardly be applied to data management for big data. This paper discusses the problems and challenges for error detection, error repair and query processing of dirty data in big data management, and identifies intractability, mixed errors and the lack of knowledge as three new challenges to data quality management. The progress of big data quality management in these three aspects is reviewed and open problems for future research are proposed.

data quality  /  big data  /  data cleaning