Testing text() nodes vs string values in XPath

Question

I have a node as follows:

<span class="portal-text-medium">Office Hours</span>

For the XPath I use

//span[text()='Office Hours']

which should work, but it never does. I can use *contains(text(),'Office Hours')]* but that won't find an exact match and I have to verify there is no "*". This is not the only time it hasn't worked for me. I have seen it work before so I don't know what is wrong. Any idea?

Yes I can, and do, use starts-with but it is not quite the same.

I have the page open in Chrome, and I do an inspect and then ctrl-f which opens a box in the lower left corner into which I can type xpaths. If the xpath matches, then the number of matches is shown and the first one is highlighted in yellow. Actually even if no match is found the number of matches is shown, and is shown as 0. When I use this xpath 0 matches are shown. When I use contains() or starts-with(), 14 or so matches are shown. — Tony
– Tony, Commented Jan 4, 2016 at 15:03

kjhughes · Accepted Answer · 2019-03-04 14:04:01Z

60

XPath `text() =` is different than XPath `. =`

(Matching text nodes is different than matching string values)

The following XPaths are not the same...

//span[text() = 'Office Hours']

Says:

Select the span elements that have an immediate child text node equal to 'Office Hours`.
//span[. = 'Office Hours']

Says:

Select the span elements whose string value is equal to 'Office Hours`.

In short, for element nodes:

The string-value of an element node is the concatenation of the string-values of all text node descendants of the element node in document order.

Examples

The following span elements would match only #1:

Office Hours 8:00-10:00
My Office Hours

The following span elements would match only #2:

Office Hours
Office Hours

The following span element would match both #1 and #2:

Office Hours

edited Mar 4, 2019 at 14:04

answered Jan 4, 2016 at 16:24

kjhughes

113k31 gold badges198 silver badges276 bronze badges

Sign up to request clarification or add additional context in comments.

7 Comments

Tony Over a year ago

So for using a ".", parts of the string can be at different depths, as long as when appended together it comes out to the value you are searching? I think sometimes when I encountered that problem I should have tried I should have used a . because maybe there were some extra levels I did not know about.

kjhughes Over a year ago

You got it. More precisely per the XPath recommendation: The string-value of an element node is the concatenation of the string-values of all text node descendants of the element node in document order. Answer updated. Thanks.

Jeyekomon Over a year ago

A relevant quote from the XPath Specification: "If one object to be compared is a node-set and the other is a string, then the comparison will be true if and only if there is a node in the node-set such that the result of performing the comparison on the string-value of the node and the other string is true."

ahdung Jul 9 at 3:26

so how to only matchxxx, not xxx xxx?

kjhughes Jul 9 at 16:07

Clearly by testing the string value of the span because the string value of the xxx is xxx while the string value of xxx xxx is xxxxxx. If you wanted to select all span elements in the document with a string value of xxx, you'd use //span[.='xxx'].

ahdung Jul 10 at 3:08

Thank you, but [.='xxx'] also match x xx right? I just want exactly match xxx

kjhughes Jul 10 at 16:51

Correct. To match xxx but not x xx (or xxx or xxx, etc), use //span[.='xxx' and not(*)] to say, "Select all span elements that have a string value of xxx and no element children."

Collectives™ on Stack Overflow

Testing text() nodes vs string values in XPath

1 Answer 1

XPath `text() =` is different than XPath `. =`

Examples

7 Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

XPath text() = is different than XPath . =

Examples

7 Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related

XPath `text() =` is different than XPath `. =`