Abstract: A method for representing, manipulating and enumerating a node set in an XML document can create and manipulate a node set in time that is typically constant or, at worst, logarithmic in the size of the underlying document, and can enumerate a node set in time that is substantially linear in the size of the node set. The method comprises steps of deriving a high-level representation of a node set. This step need not include an enumeration of the nodes of the set. The method may then manipulate the high-level representation of a node set in order to effect transformations on the node set. The high-level representation can then be analyzed in order to determine how to most efficiently process the nodes of the node set. The node set can then be efficiently enumerated, in document order, by taking advantage of the results of the previous analysis step.