Abstract: Online XML data are frequently generated in application of Web services It is feasible to employ actual XML processing methods and techniques to process online XML Most of XML storage and query strategies are based on some numbering scheme, and it is necessary to use extended numbering scheme to sustain updating well However, few studies have been reported on the issue of loading for online XML documents based on extended numbering Proposed in this paper are a new numbering scheme and a loading algorithm for online XML documents The approach makes use of statistical information about character of loaded documents with the same schema in order to finish loading in one pass Experiments show the effectiveness of these algorithms