Why does dom.firstChild.firstChild.nodeValue print the text inside the root tag?

library.xml

<?xml version="1.0" encoding="utf-8"?>
<library>library-text. :D
    <book isbn="1111111111">
        <title lang="en">T1 T1 T1 T1 T1</title>
        <date>2001</date>
        <author>A1 A1 A1 A1 A1</author>     
        <price>10.00</price>
    </book>
    <book isbn="2222222222">
        <title lang="en">T2 T2 T2 T2 T2</title>
        <date>2002</date>
        <author>A2 A2 A2 A2 A2</author>     
        <price>20.00</price>
    </book>
    <book isbn="3333333333">
        <title lang="en">T3 T3 T3 T3</title>
        <date>2003</date>
        <author>A3 A3 A3 A3 A3y</author>        
        <price>30.00</price>
    </book>
</library>

Python code

import xml.dom.minidom as minidom

xml_fname = "library.xml"

dom = minidom.parse(xml_fname) 

print(dom.firstChild.tagName)
print(dom.firstChild.firstChild.nodeValue)

Output

library
library-text. :D

Why does dom.firstChild.firstChild.nodeValue print the text inside the root tag?

Shouldn't it have been dom.firstChild.nodeValue?

Nodes in the DOM do not only represent elements, text values are also nodes. The first child node inside the <library> element is a text node, and it's value is the Python string 'library-text. :D\n ':

>>> dom.firstChild.firstChild
<DOM Text node "'library-te'...">
>>> dom.firstChild.firstChild.nodeValue
'library-text. :D\n    '

Note that the nodeValue property of Elements is always null (== None in Python); see the DOM level 1 definition for Node:

In cases where there is no obvious mapping of these attributes for a specific nodeType (e.g., nodeValue for an Element or attributes for a Comment), this returns null.

What node type holds what type of value for Node.nodeValue is specified in the Definition Group NodeType section.

The DOM API is a very bare bones, basic API, aimed at compatibility with a very broad range of languages, and this is especially true for the DOM Level 1 specification (the only spec that minidom supports). You generally don't want to use it, at all, if you can possibly avoid it. In Python, use a higher-level API like the ElementTree API (use the lxml library, which is a more feature-rich compatible implementation).

Using ElementTree, you deal primarily with just elements, and text is accessible via the text and tail attribute on elements.