Difference between these XPath expressions inside a Word Processing Markup Language?

Question

In Word Processing Markup Language (WordML) for Microsoft Office Word 2003, is there any difference between the following two XPath expressions:

@".//w:p[count(*/w:t) >= 0]"

and

@".//w:p"

I am confused because if the count is Zero, then what is the importance of the expression in the angular brackets? Doesn't both select the same number of nodes?

Also, does */w:t select only that are grandchildren or will it take into account the immediate children too?

Answer 1

This expression:

".//w:p[count(*/w:t) >= 0]"

selects all descendants of the context node (., whatever it is in the current state of processing) if they are elements with the qualified name w:p and only if their children (most likely w:r elements) have at least 0 w:telements.

That does not make a lot of sense of course, but this would:

".//w:p[count(*/w:t) >= 1]"

But actually, this would be sufficient:

".//w:p[descendant::w:t]"

The rationale behind it is to select w:pelements only if they contain text (which is stored in w:t elements, which in turn is stored in w:r elements ("runs").

---

On the other hand,

".//w:p"

selects all w:p elements that are descendants of the context node, regardless of whether they in turn contain a w:t descendant or not.

---

EDIT

Doesn't both select the same number of nodes?

Yes, both amount to the same, but one of them is a sensible expression whereas the other is not.

does */w:t select only that are grandchildren or will it take into account the immediate children too?

This expression will only take into account w:t elements that are grandchildren of w:p. Besides, this (i.e. w:tas an immediate child of w:p) is not allowed in the OOXML Schema.

Answer 2

You are right that the predicate [count(X) >= 0] is always true and can therefore be safely eliminated.