Re: trouble with preceding axis

On 1/4/06, Peter Flynn <pflynn@xxxxxx> wrote:
> I think the key information the OP was missing is that this means
> "a preceding element *called* orange" NOT "an element preceding
> the one called orange".

Nice summary.  Probably should have been a little clearer in my post
but this is the thrust of my post.

> preceding-sibling::pear is what is needed.

In this case, but I think what the OP was going for was actually to
find the closest node (not clear if it needed to be a sibling or not)
preceding orange.  The others hit it on the head with preceding::*[1]
(or preceding-sibling::*[1] which is more likely what is wanted).

> Peel me a sibling.

*lol*

Jon Gorman