Pugixml get full string containing whitespace

[adockendorf]

I'm trying to read an XML file with the structure

<name>search_name</name>
<Polygon>
	<outerBoundaryIs>
		<LinearRing>
			<coordinates>
				-25.0,30.0,0 -26.0,30.4,0 
			</coordinates>
		</LinearRing>
	</outerBoundaryIs>
</Polygon>

The coordinate string is variable length with each tuple separated by a space. Pugixml correctly finds the node tags in the larger file, but when getting the coordinate text, it will only read the first value.

Snippet of how it is being loaded/parsed:

    pugi::xml_document doc;
    pugi::xml_parse_result result = doc.load_file("file.xml");

    pugi::xpath_node_set name_nodes = doc.select_nodes("//name");
    for(const auto& xpath_node : name_nodes)
    {
        auto name_node = xpath_node.node();
        if(name_node.text().as_string() == "search_name"))
        {
            auto path_node = name_node.parent().select_node("//coordinates");
            if(path_node)
            {
                // Prints "-25.0,30.0,0"
                std::cout << "Coordinate string: " << path_node.node().text().as_string() << std::endl;
            }
        }
    }

Is this a limitation of pugixml, or is there a different way it can be parsed to preserve whitespace? I need to get the entire coordinate string, including the spaces.

[zeux]

Wild guess: second select_node is finding a node that does not correspond to your object that's currently being processed. This is a common XPath gotcha, "//foo" means "go to document root and find all nodes with the name 'foo'". You need auto path_node = name_node.parent().select_node(".//coordinates"); instead.

Otherwise, pugixml should preserve whitespace of text nodes in this case.

[adockendorf]

That was it thank you!