I'm using ADO.net (typed) datasets and serializing the data in the dataset and persisting them in xml format to a file using the .writexml method.

The problem is, I would like to incrementally add more data into the xml file without having to read the whole file again and doing a merge with the dataset containing the new data I want to add into the file.

Reading the whole file and merging is ok when it is still small but once it grows a bit larger, the performance is unacceptable?

Is there anyway to append xml data into an xml file without having to read the whole file using the dataset?

If there is no way to use the dataset then what can I use to achieve similar results?