>I do NOT want to get additionally: ><subscriberT> ><name>Name2</name> ></subscriberT> >as now, because this fragment is not valid with xml schema
The concept of taking only those portions of a document which match the schema is VERY far outside the capabilities of normal parsing. Among other things, there's the question of exactly what part of the document is the error, which requires much deeper understanding of the semantics of your particular document than anything in schemas or XSLT. The best suggestion I can give you is to parse the document without validation (so it is parsed successfully whether valid or not), then code all the validation _and_ filtering logic explicitly into your stylesheet logic. If you know only one subscriber child is expected, use [1] in your patterns to select it and ignore any others. Note that this also means your stylesheet will have to reinvent all the logic associated with default values. It would probably make more sense to insist that the source document's author repair it so it conforms to the schema. If you expect it to match a specific schema, and it doesn't, it's _wrong_ and you should not be obligated to accept it. That's the whole point of writing a schema in the first place: it establishes a specific "contract" for exactly what documents you intend to accept.
