Question

I have an XML structure similar to this:

<Header>
    <ElementA>
        <ElementB>
            <ElementC/>
            <ElementC/>
        </ElementB>
        <ElementB/>
    </ElementA>
</Header>

where the <ElementB> may have a sequence of <ElementC>, or may have none.

I can select <ElementA> nodes which have two <ElementB> by /Header/ElementA/ElementB/following-sibling::ElementB. I can select <ElementA> nodes which contain an <ElementB> node which contains an <ElementC> using /Header/ElementA/ElementB[ElementC].

But how do I select <ElementA> nodes which contain an <ElementB> which contains an <ElementC> followed by another <ElementB> containing another <ElementC>. Something like this:

<Header>
    <ElementA>
        <ElementB>
            <ElementC/>
            <ElementC/>
        </ElementB>
        <ElementB>
            <ElementC/>
            <ElementC/>
        </ElementB>
    </ElementA>
</Header>

Answer 1

Note that you don’t have to explicitly search the following-sibling axis. These two patterns will return the same result:

foo[bar/following-sibling::bar]

and

foo[bar[2]]

So your pattern can be as simple as:

/Header/ElementA[ElementB[ElementC][2]]

which will find ElementA elements that have two or more ElementB children, each of which has an ElementC child.