Reputation: 2529
I have an XPath query which is doing what I want, namely selecting a union of 'surname' & 'given-names' with the given predicates (It's actually either/or that I need, but the union works fine):
/header/authors/surname[./text() and @id='1']
|
/header/authors/given-names[./text() and @id='1']
However this seems overly long-winded to me and I feel that it should be possible to do something more succinct such as:
/header/authors/(surname|given-names)[./text() and @id='1']
... but this version is not valid XPath.
Can anyone tell me of a neater way of writing the original XPath expression which doesn't require the full path to be written twice?
Thanks
Richard
Upvotes: 11
Views: 17300
Reputation: 28004
In addition to Alejandro's and Seva's good answers, another alternative is
(/header/authors/surname | /header/authors/given-names)[text() and @id='1']
This is valid in XPath 1.0. It still is a bit redundant, but is more succinct than the long-winded version in your question.
A further alternative would be to declare a variable for surname elements and a variable for given-name elements, and then
($surnames | $given-names)[text() and @id='1']
That would probably be longer in total, but easier to read than the above.
Upvotes: 3
Reputation:
First, this is valid XPath 2.0:
/header/authors/(surname|given-names)[./text() and @id='1']
Second, this XPath 1.0:
/header/authors/*[self::surname|self::given-names][text()][@id='1']
Upvotes: 27
Reputation: 61398
How about
/header/authors/*[(name() = "surname" or name()="given-names") and ./text() and @id='1']
Upvotes: 4