Forums » Help »
Problem with FilterIterator
Added by Anonymous almost 18 years ago
Legacy ID: #3821645
Legacy Poster: Jeroen Bulters (jeroenbulters)
I am trying to write an XPath extension function and I get the following error:
java.lang.ClassCastException: net.sf.saxon.tinytree.TinyAttributeImpl at
net.sf.saxon.expr.GeneralComparison.effectiveBooleanValue(GeneralComparison.java:454) at
net.sf.saxon.expr.FilterIterator$NonNumeric.matches(FilterIterator.java:161) at
net.sf.saxon.expr.FilterIterator.getNextMatchingItem(FilterIterator.java:64) at
net.sf.saxon.expr.FilterIterator.next(FilterIterator.java:44) at
net.sf.saxon.value.SequenceExtent.<init>(SequenceExtent.java:105) at
net.sf.saxon.functions.Closure.iterate(Closure.java:129) at
net.sf.saxon.expr.PathExpression.iterate(PathExpression.java:767) at
net.sf.saxon.expr.PathExpression.iterate(PathExpression.java:767) at
net.sf.saxon.sort.DocumentSorter.iterate(DocumentSorter.java:73) at
net.sf.saxon.expr.Atomizer.iterate(Atomizer.java:96) at
net.sf.saxon.expr.ComputedExpression.process(ComputedExpression.java:600) at
net.sf.saxon.expr.ForExpression.process(ForExpression.java:540) at
net.sf.saxon.query.XQueryExpression.run(XQueryExpression.java:312) at
net.sf.saxon.Query.doQuery(Query.java:449) at net.sf.saxon.Query.main(Query.java:81)
Fatal error during transformation: net.sf.saxon.tinytree.TinyAttributeImpl This is the
code executed: NodeInfo localnode = (NodeInfo)argument[0].evaluateItem(c);
System.out.println("Document root: " + localnode.getDocumentRoot());
System.out.println("LocalNode: " + localnode);
System.out.println("LocalIter: " + c.getCurrentIterator());
StandaloneContext sc = new StandaloneContext(localnode.getDocumentRoot()); String path =
(String)argument[1].evaluateAsString(c); Expression pathexpr = ExpressionTool.make(path,
sc, 0, -1, 0); System.out.println("Path: " + path);
System.out.println("PathExpr: " + pathexpr); pathexpr.display(10,
System.out, c.getConfiguration()); XPathContextMajor newctx = new
XPathContextMajor(localnode, c.getConfiguration());
newctx.openStackFrame(c.getStackFrame().getStackFrameMap());
System.out.println("ContextItem: " +
((NodeInfo)newctx.getContextItem()).getDisplayName());
System.out.println("ContextSize: " + newctx.getLast());
SequenceIterator result = pathexpr.iterate(newctx);
System.out.println("Evaluated: " + result); The error occurs when
calling the pathexpr.iterator(<XPathContext>) method (obvious) . Does
anyone knows what I'm doing wrong?
Please register to reply
Legacy ID: #3821846
Legacy Poster: Michael Kay (mhkay)
You're using some very low-level interfaces here, so you need to be prepared to do
some low-level debugging as well! I think the basic omission is that you have failed to
run the typeCheck() method on the expression. This phase is mandatory, because it takes
care of adding checks and conversions. In this particular case I think the failure
arises because a node used in a comparison hasn't been atomized; it's the typeCheck
phase that adds the atomizer to the expression tree. There's also an optimize() phase
which in principle is optional, but I wouldn't guarantee that as Saxon always calls it.
If your expression uses variables or variable references then you also need to call
allocateSlots(). Use the source code of saxon:evaluate() to guide you. Alternatively,
consider using a higher-level API such as the JAXP API which was explicitly designed as
an external software interface.
Legacy ID: #3821921
Legacy Poster: Jeroen Bulters (jeroenbulters)
I don't think a high-level API is the solution. I'm implementing a Transitive
Closure function which is going to be benchmarked against traditional XPath methods of
determining the closure of an expression (and a Node(Set)). Thanks for the
pointers.
Legacy ID: #3821927
Legacy Poster: Jeroen Bulters (jeroenbulters)
Thanks, I think I've got it all solved now. Adding the
typeCheck(<ctx>, Type.ITEM_TYPE) worked out fine.. Now hope it passess all
the TestCases ;-)
Please register to reply
.