XPath - Difference between node() and text()

text() and node() are node tests, in XPath terminology (compare).

Node tests operate on a set (on an axis, to be exact) of nodes and return the ones that are of a certain type. When no axis is mentioned, the child axis is assumed by default.

There are all kinds of node tests:

node() matches any node (the least specific node test of them all)

text() matches text nodes only

comment() matches comment nodes

* matches any element node

foo matches any element node named "foo"

processing-instruction() matches PI nodes (they look like ).

Side note: The * also matches attribute nodes, but only along the attribute axis. @* is a shorthand for attribute::*. Attributes are not part of the child axis, that's why a normal * does not select them.

This XML document:

<produce>
    <item>apple</item>
    <item>banana</item>
    <item>pepper</item>
</produce>

represents the following DOM (simplified):

root node
   element node (name="produce")
      text node (value="\n    ")
      element node (name="item")
         text node (value="apple")
      text node (value="\n    ")
      element node (name="item")
         text node (value="banana")
      text node (value="\n    ")
      element node (name="item")
         text node (value="pepper")
      text node (value="\n")

So with XPath:

/ selects the root node

/produce selects a child element of the root node if it has the name "produce" (This is called the document element; it represents the document itself. Document element and root node are often confused, but they are not the same thing.)

/produce/node() selects any type of child node beneath /produce/ (i.e. all 7 children)

/produce/text() selects the 4 (!) whitespace-only text nodes

/produce/item[1] selects the first child element named "item"

/produce/item[1]/text() selects all child text nodes (there's only one - "apple" - in this case)

And so on.

So, your questions

"Select the text of all items under produce" /produce/item/text() (3 nodes selected)

"Select all the manager nodes in all departments" //department/manager (1 node selected)

Notes

The default axis in XPath is the child axis. You can change the axis by prefixing a different axis name. For example: //item/ancestor::produce

Element nodes have text values. When you evaluate an element node, its textual contents will be returned. In case of this example, /produce/item[1]/text() and string(/produce/item[1]) will be the same.

Also see this answer where I outline the individual parts of an XPath expression graphically.

---

For me it was a big difference when I faced this scenario (here my story:)

<?xml version="1.0" encoding="UTF-8"?>
<sentence id="S1.6">When U937 cells were infected with HIV-1, 
        
    <xcope id="X1.6.3">
        <cue ref="X1.6.3" type="negation">no</cue> 
                        
                        induction of NF-KB factor was detected
        
    </xcope>
                    
, whereas high level of progeny virions was produced, 
        
    <xcope id="X1.6.2">
        <cue ref="X1.6.2" type="speculation">suggesting</cue> that this factor was 
        <xcope id="X1.6.1">
            <cue ref="X1.6.1" type="negation">not</cue> required for viral replication
        </xcope>
    </xcope>.

</sentence>

I needed to extract text between tags and aggregate (by concat) the text including in innner tags.

/node() did the job, while /text() made half job

/text() only returned text not included in inner tags, because inner tags are not "text nodes". You may think, "just extract text included in the inner tags in an additional xpath", however, it becomes challenging to sort the text in this original order because you dont know where to place the aggregated text from the inner tags!because you dont know where to place the aggregated text from the inner nodes.

When U937 cells were infected with HIV-1, no induction of NF-KB factor was detected , whereas high level of progeny virions was produced, suggesting that this factor was not required for viral replication .

Finally, /node() did exactly what I wanted, because it gets the text from inner tags too.

---