User Controllable Data Grouping for Business Document Translation

Business document translation is a critical business activity that is essential for business process integration. To achieve interoperability among enterprise applications, common document formats or standards must be followed across business entities, e.g., the legacy electronic data interchange (EDI) and the popular extensible Markup Language (XML). Based on the document object model (DOM), both source and target documents can be visualized as structural trees. Naturally, document translation involves a tree traversal process (for source) and a tree creation process (for target). However, data grouping problem occurs when there are multiple items of the same type (or XML tag) and there is a need to group items differently. The normal document translation process may need to traverse the source DOM trees multiple times due to the ambiguities associated with the process. In this paper, we propose a document translation mechanism that performs a tree traversal over the source tree structure in one pass and automatically creates a target tree structure based upon user-defined rules. For certain grouping options that may cause ambiguity during the translation process, our approach requires just an additional pass over the target tree structure.