Tirgul OOP No.3

1
Tirgul OOP No.3

• Iterators

2
What is an Iterator?

• An object that provides a way to access elements
  of an aggregate object sequentially, without
  exposing its internal structure

3
Example

• Suppose we have a myList object, which implements
  the List interface
• public void remove(Object item)
• public void add(Object item)
• public Object elementAt(int index)
• We would like to print all of the elements in the
  Cartesian product (myList X mylist)

4
Naive Solution - 1

• Add a function to the list interface which
  performs the task.
• Why is that bad ?
• What about myList X myList X myList ?
• What if we change the list implementation ?

5
Naive Solution - 2
void printAllPairs(IntList list) for(int i0
iltlist.size() i) int e1
list.elementAt(i) for( int j0
jltlist.size() j) int e2
list.elementAt(j) System.out.println
(lte1,e2gt)
Pros and cons?
6
Outline

• Definition
• Motivation
• Basic Implementation and usage
• Implementation issues

7
Declaring an Iterator

• The Java iterator interface
• public class Iterator
• public Iterator()
• public boolean hasNext()
• public Object next()
• Create a ListIterator implementation for List
• Add the following to the List interface
• public Iterator iterator()

8
ImplementationAssuming Linked List
public class LinkedList . . . . . . private
class LinkedListIterator . . . Node
_position . . . LinkedListIterator(Node
head) _position head . . .
9
ImplementationAssuming Linked List
public Object next() . . .
// Check validity Object ret
_position._data _position
_position._next return ret public
boolean hasNext() return
_position ! null public Iterator iterator()
return new LinkedListIterator(_head)
10
Using an Iterator
void printAllPairs(List list) Iterator i1
for (i1 list.iterator() i1.hasNext() )
Iterator i2 Object e1 i1.next() for
(i2 list.iterator() i2.hasNext())
Object e2 i2.next()
System.out.println(lte1,e2gt)

11
Outline

• Definition
• Motivation
• Basic Implementation and usage
• Implementation issues
• Who defines the traversal algorithm?
• Using auxiliary data structures
• Access rights of an iterator
• Iterators for composites

12
Who defines the traversal algorithm?

• Options
• The aggregate object
• Next operation belongs to the aggregate object
• The iterator merely points to the current
  position
• The iterator
• Better separation from the point of view of the
  client
• Might require privileged access rights

13
Using an Auxiliary Data Structure
public class List . . . . . . public
Iterator filtIterator(Object fltr) List
auxList new List Iterator i for (i
iterator() i.hasNext() ) Object
curItem i.next() if (!curItem.equals(fltr
)) auxList.add(curItem)
return new FiltIterator(auxList) . . .
14
Iterator Defined Traversal - 1
public class List . . . private class
FilteredIterator . . . Node _current
Object _filter public FilteredIterator
(Node head, Object filter)
_current head _filter filter
advanceToNext() public boolean
hasNext() return (_current ! null)

Note the access rights of the iterator
15
Iterator Defined Traversal - 2
public Object next()
advanceToNext() . . . //
Check validity return _current
void advanceToNext() boolean
flag true while(flag _current !
null) if( _current.data.equals(_filter)
) _current _current.next
else flag false

16
Designing Iterators for Trees

• Options
• Store a path to keep track of the current
  positions
• Use an auxiliary list
• Enhance the composite data structure
• Add access to children, parents and siblings
• Use an internal iterator