When using the "XML reader" node to "read" an XML file, I've found that the node adds "extra blanks" to the resulting file. For example, when "reading" the very small file "license.xml" (Attached below), I count four blank spaces before the less-than sign that marks the begininng of the tag <NAME> (or the tag <ID>, <KEY>, etc):.
What you see in KNIME is always a pretty-printed version of the XML file. If there is white-space before an opening tag or not is completely irrelevant for the XML tree itself.
I understand what you are saying. But I'm asking because I need to compare entire XML files between what KNIME produces via the XML reader, and a database, where those same XML files are stored (as CLOBs). When I use PipelinePilot to read an XML file, the ouput respects the original white spaces in the XML file, whereas KNIME adds additional white spaces (Which affects a full file comparison later on). That's why I thought the KNIME node could be tweaked to prevent the addition of extra white spaces, so it would produce the same file (including whitespaces) that you get when you open it in a text editor (e.g. text wrangler) or via PP, as mentioned.
Comparing XML files character-wise is a bad idea, exactly for such reasons. To make a correct comparison you either have to compare the structure (you can use the Table Difference Checker node from the KNIME Testing Framework extension for this) or you process both XML files with KNIME so that they are pretty-printed the same way.