Data Warehouse is a logical stockroom which accumulates and maintains enormous volume of data. It is very important for the enterprise that needs to analyze data obtained from heterogeneous sources to take tactical decisions. Any enterprise that wants to expand, survive and beat out the competition must have control over data. Any error or inconsistency in the data may lead to improper decision which may cause immense losses to enterprise. Data Warehouse testing is carried out to eliminate the errors and inconsistencies that arise due to data being collected from desperate sources in different formats. Data Warehouse testing is too expensive and time consuming practice as exhaustive testing is not possible. Therefore the concept of Data Mart comes into existence. Data Mart is a specialized subset of Data Warehouse which fulfills the data requirement of a specific group. Testing of a Data Mart is much easier and manageable process as compared to testing of a Data Warehouse. In order to test the data mart, there are a number of strategies that allow us to select a set of test cases and test data which are very effective in detecting errors. In our paper we have discussed black box and white box testing techniques concerning Data Mart in brief. We have explored black box techniques to select test cases as they have systematic approach to uncover a great number of errors. We have also proposed to design the test cases using Boundary Value Analysis, Equivalence Class Partitioning and the combination of both techniques. The test cases designed using these techniques are very successful to detect previously undetected errors or faults in a Data Mart. These are also proficient for testing the worst case and robustness of Data Mart.

Data Warehouse Data Mart Data Mart testing Black box testing technique White box testing technique Equivalence class partitioning Boundary value analysis