Class OffsetOrderVisitor<T extends QueryTreeNode>

java.lang.Object
org.apache.derby.impl.sql.compile.OffsetOrderVisitor<T>
Type Parameters:
T - the type of nodes to collect
All Implemented Interfaces:
Visitor

class OffsetOrderVisitor<T extends QueryTreeNode> extends Object implements Visitor
Get all nodes of a certain type in a query tree, and return them in the order in which they appear in the original SQL text. This visitor is useful when rewriting SQL queries by replacing certain tokens in the original query.
  • Field Details

    • COMPARATOR

      private static final Comparator<QueryTreeNode> COMPARATOR
      Comparator that orders nodes by ascending begin offset.
    • nodeClass

      private final Class<T extends QueryTreeNode> nodeClass
    • nodes

      private final TreeSet<T extends QueryTreeNode> nodes
    • lowOffset

      private final int lowOffset
    • highOffset

      private final int highOffset
  • Constructor Details

    • OffsetOrderVisitor

      OffsetOrderVisitor(Class<T> nodeClass, int low, int high)
      Create a new OffsetOrderVisitor that collects nodes of the specified type. The nodes must have begin offset and end offset in the range given by the low and high parameters.
      Parameters:
      nodeClass - the type of nodes to collect
      low - the lowest begin offset to accept (inclusive)
      high - the highest end offset to accept (exclusive)
  • Method Details

    • visit

      public Visitable visit(Visitable node) throws StandardException
      Description copied from interface: Visitor
      This is the default visit operation on a QueryTreeNode. It just returns the node. This will typically suffice as the default visit operation for most visitors unless the visitor needs to count the number of nodes visited or something like that.

      Visitors will overload this method by implementing a version with a signature that matches a specific type of node. For example, if I want to do something special with aggregate nodes, then that Visitor will implement a visit(AggregateNode node) method which does the aggregate specific processing.

      Specified by:
      visit in interface Visitor
      Parameters:
      node - the node to process
      Returns:
      a query tree node. Often times this is the same node that was passed in, but Visitors that replace nodes with other nodes will use this to return the new replacement node.
      Throws:
      StandardException - may be throw an error as needed by the visitor (i.e. may be a normal error if a particular node is found, e.g. if checking a group by, we don't expect to find any ColumnReferences that aren't under an AggregateNode -- the easiest thing to do is just throw an error when we find the questionable node).
    • visitChildrenFirst

      public boolean visitChildrenFirst(Visitable node)
      Description copied from interface: Visitor
      Method that is called to see if visit() should be called on the children of node before it is called on node itself. If this method always returns true, the visitor will walk the tree bottom-up. If it always returns false, the tree is visited top-down.
      Specified by:
      visitChildrenFirst in interface Visitor
      Parameters:
      node - the top node of a sub-tree about to be visited
      Returns:
      true if node's children should be visited before node, false otherwise
    • stopTraversal

      public boolean stopTraversal()
      Description copied from interface: Visitor
      Method that is called to see if query tree traversal should be stopped before visiting all nodes. Useful for short circuiting traversal if we already know we are done.
      Specified by:
      stopTraversal in interface Visitor
      Returns:
      true/false
    • skipChildren

      public boolean skipChildren(Visitable node) throws StandardException
      Description copied from interface: Visitor
      Method that is called to indicate whether we should skip all nodes below this node for traversal. Useful if we want to effectively ignore/prune all branches under a particular node.

      Differs from stopTraversal() in that it only affects subtrees, rather than the entire traversal.

      Specified by:
      skipChildren in interface Visitor
      Parameters:
      node - the node to process
      Returns:
      true/false
      Throws:
      StandardException
    • getNodes

      SortedSet<T> getNodes()