getChildNodes() 返回 Text 而不是 NodeList

Question

I'm trying to do XML parsing in MATLAB, which uses JavaSE package org.w3c.dom

I have a xml structure like

 <?xml version="1.0" encoding="utf-8"?> <tada> <Filters> <filter id="1" name="name1" renderType="type1"> <bool name="par1" value="false"/> <bool name="par2" value="false"/> <int name="par3" value="200"/> <int name="par4" value="1"/> <float name="par5" value="2"/> <bool name="par6" value="false"/> <int name="par7" value="1"/> </filter> <filter id="6" name="name2" renderType="type2"> <bool name="par1" value="false"/> <bool name="par2" value="false"/> <int name="par3" value="200"/> <int name="par4" value="1"/> <float name="par5" value="2"/> <bool name="par6" value="false"/> <int name="par7" value="1"/> <int name="par8" value="1"/> <array name="par9" row="2" col="2" > 5.0162 0.96182 0.22294 0.96356 </array> </filter> </Filters> </tada>

I'm trying to get access for, let say attribute renderType of element filter with name="name2"

I'm trying to do the following

 config = xmlread(xmlFileName); section = config.getDocumentElement().getElementsByTagName('Filters').item(0); filters = section.getChildNodes(); // filters = section.getElementsByTagName('filter'); for i = 0:filters.getLength()-1 filter = filters.item(i); if strcmp(filter.getAttribute('name'), 'name2') filter.setAttribute('renderType', 'graph2D'); end end

I expect that filters should be a NodeList (as documentation says http://docs.oracle.com/javase/6/docs/api/org/w3c/dom/package-summary.html?browser=F1help ), but it's returned as a Text (which contains only white spaces).

If I write

 filters = section.getElementsByTagName('filter');

then I get desired NodeList , but the problem is that this way is unsuitable. Because I need to select nodes by the value of name property.

Could you help me? What I do wrong?

Thank you for your help.

Answer 1

Your XML document includes whitespace between the elements. This is considered significant by the XML parser, and when you use the DOM method getChildNodes() it includes both the child elements and the child text nodes containing the whitespace. When processing the child nodes, if you want to ignore the whitespace text nodes you must do so explicitly.