ABSTRACT
The conflict between real-time data import and real-time data query will arise in the real-time data warehouse [1]. One of the most difficult parts in the construction of any data warehouse is to import data from different business systems to ECCL. More additional difficulties will be increased if the process becomes real-time. Almost all systems are in operation in batch mode. In these systems, the available data is assumed to exist in a place in a kind of extracted file within a certain definite time schedule. Then, the systems transform and clean the data, and then import it to the data warehouse [2]. The processing exerts a significant influence on the nonresponse period of the data warehouse, for no users can access the data warehouse in the process of import. But with the continuous and real-time import of data, the system could not have any nonresponse periods. The most cumbersome retrieval time period for the data warehouse may be consistent with the time when the data is imported most frequently. Therefore, a fundamental contradiction has been produced between the system and this requirement for continuous update without nonresponse periods [3].
