FLWOR Expression

[Philippe Michiels]

> //chunklist/chunk/par/s/tok[orth/text() = "se" and
> contains(disamb/ctag/text(), "KS")]

Do you mean something like this:

for $q in
	for $p in
		...
		for $z in
			for $y in
				for $x in $root
				return $x/descendant::chunklist
			return $y/child::chunk
		return $z/child::par
		...
	return $p/child::stok
where
	<<predicate expression here >>
return $q

Some remarks: XPath results are sorted according to document order and contain 
no duplicates. Decomposing them into loops does give the same result up to 
ordering and duplicates.

Best,
Ph