Class FocusTrackingIterator

  • All Implemented Interfaces:
    java.io.Closeable, java.lang.AutoCloseable, LastPositionFinder, FocusIterator, SequenceIterator, GroundedIterator, LookaheadIterator
    Direct Known Subclasses:
    MultithreadedFocusTrackingIterator

    public class FocusTrackingIterator
    extends java.lang.Object
    implements FocusIterator, LookaheadIterator, GroundedIterator, LastPositionFinder
    An iterator that maintains the values of position() and current(), as a wrapper over an iterator which does not maintain these values itself.

    Note that when a FocusTrackingIterator is used to wrap a SequenceIterator in order to track the values of position() and current(), it is important to ensure (a) that the SequenceIterator is initially positioned at the start of the sequence, and (b) that all calls on next() to advance the iterator are directed at the FocusTrackingIterator, and not at the wrapped SequenceIterator.

    Since:
    9.6
    • Constructor Detail

      • FocusTrackingIterator

        public FocusTrackingIterator​(SequenceIterator base)
    • Method Detail

      • getUnderlyingIterator

        public SequenceIterator getUnderlyingIterator()
        Get the underlying iterator
        Returns:
        the iterator underlying this FocusIterator
      • next

        public Item next()
        Get the next item in the sequence. This method changes the state of the iterator, in particular it affects the result of subsequent calls of position() and current().
        Specified by:
        next in interface SequenceIterator
        Returns:
        the next item, or null if there are no more items. Once a call on next() has returned null, no further calls should be made. The preferred action for an iterator if subsequent calls on next() are made is to return null again, and all implementations within Saxon follow this rule.
        Since:
        8.4
      • current

        public Item current()
        Get the current value in the sequence (the one returned by the most recent call on next()). This will be null before the first call of next(). This method does not change the state of the iterator.
        Specified by:
        current in interface FocusIterator
        Returns:
        the current item, the one most recently returned by a call on next(). Returns null if next() has not been called, or if the end of the sequence has been reached.
        Since:
        8.4
      • position

        public int position()
        Get the current position. This will usually be zero before the first call on next(), otherwise it will be the number of times that next() has been called. Once next() has returned null, the preferred action is for subsequent calls on position() to return -1, but not all existing implementations follow this practice. (In particular, the EmptyIterator is stateless, and always returns 0 as the value of position(), whether or not next() has been called.)

        This method does not change the state of the iterator.

        Specified by:
        position in interface FocusIterator
        Returns:
        the current position, the position of the item returned by the most recent call of next(). This is 1 after next() has been successfully called once, 2 after it has been called twice, and so on. If next() has never been called, the method returns zero. If the end of the sequence has been reached, the value returned will always be <= 0; the preferred value is -1.
        Since:
        8.4
      • getLength

        public int getLength()
        Get the position of the last item in the sequence. The method is stateless in its external effect: that is, it does not change the values returned by position(), next(), hasNext(), etc. However, it creates a new base iterator which means that the result of calling getUnderlyingIterator() may change.
        Specified by:
        getLength in interface FocusIterator
        Specified by:
        getLength in interface LastPositionFinder
        Returns:
        the position of the last item
        Throws:
        UncheckedXPathException - if a failure occurs reading the sequence
      • supportsGetLength

        public boolean supportsGetLength()
        Ask whether this iterator supports use of the getLength() method. This method should always be called before calling getLength(), because an iterator that implements this interface may support use of getLength() in some situations and not in others
        Specified by:
        supportsGetLength in interface LastPositionFinder
        Returns:
        true if the getLength() method can be called to determine the length of the underlying sequence.
      • supportsHasNext

        public boolean supportsHasNext()
        Description copied from interface: LookaheadIterator
        Ask whether the hasNext() method can be called. This method must be called before calling hasNext(), because some iterators implement this interface, but only support look-ahead under particular circumstances (this is usually because they delegate to another iterator)
        Specified by:
        supportsHasNext in interface LookaheadIterator
        Returns:
        true if the LookaheadIterator.hasNext() method is available
      • hasNext

        public boolean hasNext()
        Determine whether there are more items to come. Note that this operation is stateless and it is not necessary (or usual) to call it before calling next(). It is used only when there is an explicit need to tell if we are at the last element.

        This method must not be called unless the method supportsHasNext() returns true.

        Specified by:
        hasNext in interface LookaheadIterator
        Returns:
        true if there are more items in the sequence
        Throws:
        java.lang.ClassCastException - if the base iterator does not support lookahead processing
      • materialize

        public GroundedValue materialize()
        Return a GroundedValue containing all the items in the sequence returned by this SequenceIterator. This should be an "in-memory" value, not a Closure.
        Specified by:
        materialize in interface GroundedIterator
        Returns:
        the corresponding Value
        Throws:
        UncheckedXPathException - in the cases of subclasses (such as the iterator over a MemoClosure) which cause evaluation of expressions while materializing the value.
      • getResidue

        public GroundedValue getResidue()
        Return a GroundedValue containing all the remaining items in the sequence returned by this SequenceIterator, starting at the current position. This should be an "in-memory" value, not a Closure.
        Specified by:
        getResidue in interface GroundedIterator
        Returns:
        the corresponding Value
        Throws:
        UncheckedXPathException - in the cases of subclasses (such as the iterator over a MemoClosure) which cause evaluation of expressions while materializing the value.
      • close

        public void close()
        Close the iterator. This indicates to the supplier of the data that the client does not require any more items to be delivered by the iterator. This may enable the supplier to release resources. After calling close(), no further calls on the iterator should be made; if further calls are made, the effect of such calls is undefined.

        (Currently, closing an iterator is important only when the data is being "pushed" in another thread. Closing the iterator terminates that thread and means that it needs to do no additional work. Indeed, failing to close the iterator may cause the push thread to hang waiting for the buffer to be emptied.)

        Specified by:
        close in interface java.lang.AutoCloseable
        Specified by:
        close in interface java.io.Closeable
        Specified by:
        close in interface SequenceIterator
        Since:
        9.1
      • getSiblingPosition

        public int getSiblingPosition​(NodeInfo node,
                                      NodeTest nodeTest,
                                      int max)
        Get the sibling position of a node: specifically, count how many preceding siblings of a node satisfy the nodetest. This method appears here because it can potentially make use of cached information. When an instruction such as xsl:apply-templates select="*" (which selects a set of sibling nodes) is used in conjunction with patterns such as match="*[position() mod 2 = 1], then calculation of the position of one node in the sequence of siblings can take advantage of the fact that the position of the immediately preceding sibling is already known.

        This optimization was suggested by the late Sebastian Rahtz, one of Saxon's earliest power users, and it is dedicated to his memory.

        Parameters:
        node - the starting node, which is assumed to satisfy the node test
        nodeTest - the node test
        max - the maximum number of nodes to be counted
        Returns:
        the number of preceding siblings that satisfy the node test, plus one, unless the number exceeds max, in which case return some number greater than or equal to max.