Given the desire to use XPath to select an XML Text Node based on its value, is there a better way than:
//*[text()="foo"]/text() I'm assuming that there's something like:
//text()[self="foo"] ...but I don't know what to put for self.
Add a comment |
2 Answers
In XPath the period or dot character . is the self, also known as the "context node".
This query…
//text()[.="foo"] …will select all text nodes with the value of "foo".
3 Comments
pguardiario
I think the * is a typo, yes?
Phrogz
@pguardiario It certainly was, yes. Thanks.
Dimitre Novatchev
Good answer, but there are some peculiarities to be considered, such as the many times unnecessary use of
// and not taking into account surrounding white space.You have generally found the right answer:
//text()[.="foo"] However, please note the following:
The XPath
//pseudo-operator should be avoided as much as possible whenever the structure of the XML document is statically known, because its evaluation causes complete subtrees (in the worst case the complete document tree) to be traversed and is very inefficient. If you know that all text nodes are children ofcand allcin the document can be selected by/a/b/c, then use:/a/b/c/text(). If you only know that all text nodes are descendants ofc, then it is still much better to use:/a/b/c//text()than//text()or even than//c//text()Quite often the text nodes contain whitespace that we tend not to notice. In such cases one needs to write:
normalize-space(.) = 'foo'rather than
. = 'foo'
answered Nov 3, 2011 at 12:58
Dimitre Novatchev
244k27 gold badges308 silver badges438 bronze badges
1 Comment
Phrogz
You get the accept mark because: you confirmed what seemed to work; you are not me; the excellent note on
normalize-space. Thanks (as always) for sharing your XPath expertise.