Generating Nested XML DTD from Non Normalized Relational Views

Document Type Definition (DTD) is very important because it contains a set of rules that specifies and captures the structure of XML (eXtensible Markup Language) document. Most of the current approaches for generating DTD from relational views do not support nested elements for DTD. Furthermore, these approaches may not generate automatically DTD from flat relational views where they depend on what is submitted by users. This paper investigates how nested XML DTD that support nested elements can be automatically generated from flat relational view. The proposed approach is based on the number of data values for each column in nested view that is generated from flat relational view. The data values for each column of nested view are counted to classify the columns into groups. Each group has the columns with the same number of values. The generated groups represent the DTD elements where the element that has a bigger number of data values is nested into the element that has a smaller number of data values. According to the experimental results, the proposed approach can reduce the storage size of generated DTD by around 31% compared to the other approaches.