In data mining, the authors are extracting data from historical knowledge and create data sets. Many hyper graph concepts and data mining algorithms are used to create datasets in tabular format which consist complex queries, joining tables and aggregated functions like min, sum, count, max etc. This task is very hard and time taking. Tabular format takes suitable input to prepare data sets. But, Existing SQL aggregations having limited capacity to prepare data sets. They go back one column per aggregated group.