ECS510

Algorithms and Data Structures in an

Object Oriented Framework
“ADSOOF” 


Java’s Collections

Framework

 Java’s Collections Framework

• A “framework” is a group of classes and interfaces
designed to work together

• Java’s Collections Framework provides library
code for collections of data

• Using library code saves you programming effort

• It means you are using a common standard, others
will be familiar with that standard, and will more
easily understand your code

• Your code will more easily link with other code
which uses the same standard

 ArrayList

• ArrayList<E> is just one class in the
Collections Framework

• It is an implementation of the interface List<E>

• List<E> is an extension of another interface,
Collection<E>

• List<E> is intended to define array-like
behaviour

• Unless you have particular need for “hands-on”
control, it’s best to program using these rather
than arrays directly

 Collection!

List! Set

! SortedSet!
ArrayList
LinkedList! HashSet

TreeSet!
 Collection<E>

Has methods which include:

!boolean add(E e)!
!boolean remove(Object o)!
!boolean contains(Object o)!
!int size()

Doesn’t fully define how they work




 Set<E>

• Defines [Link](e) to work as:

– If [Link](e) was false, it becomes true,
method returns true, [Link]() increased by 1

– If [Link](e) was true, it remains true,
method returns false, [Link]() unchanged

• Defines [Link](e) to work as:

– If [Link](e) was true, it becomes false,
method returns true, [Link]() decreased by 1

– If [Link](e) was false, it remains false,
method returns false, [Link]() unchanged

• This is mathematical set behaviour

 List<E>

• Adds methods which work on items being in a
position in the collection:

– E get(i)

– E set(i,e) (returns replaced element)

– int indexOf(e)

• Defines [Link](e) to work by adding e to end
of list, returns true

• Defines [Link](e) to work by removing
lowest indexed occurrence of e, returns false if
no occurrence of e, true otherwise

 Overloaded List<E> methods


Overloading is when there are two methods of the
same name in a class or interface, List<E> has two
examples:

• [Link](i,e) adds e at position i in list, following
elements move up one position (different from
[Link](e) as that has one argument, this has
two)

• [Link](i) removes the element at position i
in list, following elements down one position
(different from the general remove method only by
argument type, int, rather than Object)

 Copy constructors


Collection classes do not have a constructor which
works as constructors for new arrays do, that is by
creating a collection of a particular size. But they do
have constructors which take another collection
object as their argument and create a new collection
containing that collection’s elements

• ArrayList(Collection<? extends E> c)!

• HashSet(Collection<? extends E> c)

• TreeSet(Collection<? extends E> c)

 Iterator<E>

• Class Collection<E> has method


Iterator<E> iterator()

• Class Iterator<E> has methods

– E next() returns an item from the collection that hasn’t
been returned by a previous next() call, throws
NoSuchElementException if there are no more

– boolean hasNext() returns true if there are further
items to be returned by a next() call, false otherwise

• You can’t use for(int i=0; i<[Link](); i++)
to go through a Set, but you can use Iterator

 For-each loop

for(Iterator<E> it=[Link](); [Link](); )!
{!
E element = [Link]();!
...!
}!
is a common pattern, Java (since Java 5) allows a shorthand
equivalent:

for(E element: coll)!
{!
...!
}!
This works for coll of type Collection<E> or any
extension, and also for coll of type E[]
Order of return for Iterator<E>

• When produced from an object which
implements List<E>, elements are returned
in their position order

• When produced from a HashSet<E> object,
elements are not returned in any obvious order

• When produced from a TreeSet<E> object,
elements are returned in their natural order (but
see later for variation of this)

 Other Collection<E> methods

• boolean addAll(Collection<? extends E> c)!
• boolean containsAll(Collection<?> c)!
• boolean removeAll(Collection<?> c)!
• boolean retainAll(Collection<?> c)!
• Object[] toArray()!
• <T> T[] toArray(T[] a)!
• boolean isEmpty()!
 Other List<E> methods

• boolean addAll(int index,

! ! ! ! !Collection<? extends E> c)!
• int lastIndexOf(Object o)!
• ListIterator<E> listIterator()!
• List<E> subList(int fromIndex, int toIndex)!
 SortedSet<E>

• SortedSet<E> introduces some extra methods
to Set<E>

– SortedSet<E> headSet(e) where
[Link](e) returns the set of all elements from s
less than e!
– SortedSet<E> tailSet(e) where
[Link](e) returns the set of all elements from s
greater than or equal to e!
• TreeSet<E> implements SortedSet<E>

HashSet<E> and TreeSet<E>

• HashSet<E> is implemented using a “hash
table”, elements stored in array, “hash function”
gives their position, O(1) access

• TreeSet<E> is implemented using an “ordered
binary tree”, linked structure, cells have two links,
follow one or other in binary search to find
element, O(log N) access

• There is not time to discuss these data structures in
more detail in this module

• See data structures text books and web sites

 Comparator<T>

• Provided with Java, it is:

!interface Comparator<T>!
!{!
! int compare(T o1, T o2)!
!}!
• If c is of type Comparator<T> and t1 and t2
are of type T, then [Link](t1,t2) returns

– A negative integer if t1 less than t2

– A positive integer if t1 greater than t2

– 0 if they are equal


in some ordering

 Writing your own Comparator

• You can write your own class which implements
Comparator<T> for T some type, but which gives an
ordering different from T’s natural order, example:

!class LengthComparer implements Comparator<String>!
!{!
! !public int compare(String str1,String str2)!
! !{!
! ! return [Link]()-[Link]();!
! !}!
!}!

This orders Strings by length rather than alphabetically

 Scrabble score Comparator

class Scrabble implements Comparator<String>!
{!
public static final int[] scores =
{1,3,3,2,1,4,2,4,1,8,5,1,3,1,1,3,10,1,1,1,1,4,4,8,4,10};!
!
public static int score(String str) {!
String str1 = [Link]();!
int score = 0;!
for(int i=0; i<[Link](); i++)!
score+=scores[[Link](i)-'A'];!
return score;!
}!
!
public int compare(String str1,String str2) {!
return score(str1)-score(str2);!
}!
}
Finding the biggest using a Comparator

public static <T> T !
biggest(Collection<T> coll, Comparator<? super T> comp)!
{!
Iterator<T> it = [Link]();!
T biggest = [Link]();!
while([Link]())!
{!
T nextItem = [Link]();!
if([Link](nextItem,biggest)>0)!
biggest = nextItem;!
}!
return biggest;!
}!
!

You can use this to find the biggest item in a collection according
to the ordering given by the Comparator argument

• If we have ArrayList<String> words



– biggest(words,new LengthComparer())



returns the longest string from words

– biggest(words,new Scrabble())!


returns the string with the highest Scrabble score!
• But rather than write our own biggest, we can use Java’s
method max in class Collections

– [Link](words,new LengthComparer())


returns the longest string from words!
— [Link](words,new Scrabble())!
! returns the string with the highest Scrabble score

• Comparator<? super T> allows us to use e.g. a
Comparator<Fruit> to find the biggest from a
Basket<Banana>



Sorting using a Comparator

• We have already seen Java’s built-in methods which sort by
“natural order”, so if ls is of type List<E> then the
method call [Link](ls) will sort the list to
which ls refers. The sorting is done “in place”.

• This means the elements of ls must have their own
compareTo method, so E should be declared as
<E extends Comparable<? super E>>!
• As an alternative, [Link](ls,comp) will
sort the list to which ls refers by the order given by a
comparator referred to by comp.

• With this, E can be any type, so does not need a bound, and
comp should be of type Comparator<? super E>!
 Anonymous Classes

• Comparator<String> c = new Comparator<String>(){!
! !public int compare(String str1,String str2)!
{!
return [Link]()-[Link]();!
}!
};!

is an alternative to

• Comparator<String> c = new LengthComparer())!

It avoids the need to have a separate class

LengthComparer
 Constructing a TreeSet using a
Comparator

• new TreeSet<Thing>() will construct a new
TreeSet<Thing> object with its contents ordered
by the natural order of Thing (as given by its
compareTo method)

• If c is of type Comparator<? super Thing> then
new TreeSet<Thing>(c) will construct a new
TreeSet<Thing> object with its contents ordered
by the compare method in c.

• The variant ordering will be observed in the
Iterator produced from the TreeSet, and
also in the way headSet etc works

 Map<K,V>

• Map<K,V> is a generic interface with two type
arguments

• It represents a collection of objects of type V
indexed by keys of type K

• K is often but not always String

• HashMap<K,V> and TreeMap<K,V> are two
classes which implement it

• TreeMap<K,V> has elements stored in order of K

• HashMap<K,V> more efficient but no ordering of
elements

 Map methods

• With m of type Map<K,V>, tag of type K and
item of type V:

– [Link](tag,item) puts item into the map with
key tag, returns previous item with key tag!
– [Link](tag) returns the last item put in with key
tag so long as it hasn’t been removed

– [Link](tag) removes the item with key tag

– [Link]() returns an object of type Set<K>
giving all the keys of items in the map

– [Link]() return an object of type
Collection<V> giving all the items in the map

 Map uses

• A collection whose elements are indexed by an
identifier (key) rather than a position

• Unlike lists, adding or deleting an item does not
change the index of other items

• Could be used as a simple database - K is student no.,
NI no., car reg, V is class representing students,
citizens, cars

• Or anywhere else where we have data as a collection
of pairs, with one part of the pair not duplicated in
other pairs e.g. count of number of occurrences of
each word in a document, K is String for word, V is
Integer for the count

 SubLists

• If str is of type String then
[Link](from,to) is a String of
the characters in str starting at position from up
to but not including position to

• Similarly, if a is of type List<E> (so
ArrayList<E> or LinkedList<E>) then
[Link](from,to) is a List<E> of the
elements in a starting at position from up to but
not including position to

• Example, a is [10,20,30,40,50,60,70], so
b=[Link](2,5) sets b to [30,40,50]!
 Views

• A view of a collection is a collection which shares
(some of) the data of the collection

• Changing the view changes the collection, and
vice-versa

• b=[Link](p1,p2) creates a view of a

• If s if of type TreeSet<E> and e is of type E,
then [Link](e) and [Link](e)
create views of s

• b=[Link](a)
creates a view of a

 SubLists are views

• Example, a is [10,20,30,40,50,60,70],
so b=[Link](2,5) sets b to
[30,40,50]!
• This means [Link](1,35) changes b to
[30,35,40,50] and changes a to
[10,20,30,35,40,50,60,70]!
• After this [Link](4,45) changes a to
[10,20,30,35,40,45,50,60,70] and
changes b to [30,35,40,45,50]

 Unmodifiable views

• b=[Link](a)
creates a view of a!
• If you attempt to call any method on b which
changes it, e.g. add or remove, you will get an
UnsupportedOperationException

• This is useful if you want another part of a
program to access your data but not to change it

• But if a call on a changes a, b will get changed in
the same way!
 Java’s Collections Framework

• Consists of interfaces, implementations and
algorithms

• Interfaces specify an Abstract Data Type, and are
arranged in a hierarchy (an interface may specialise
another interface)

• Implementations use data structures to provide
working code - different implementations have
different benefits

• Some algorithms, e.g. sort, provided as static methods

• There are various extra interfaces to aid writing
generalised code

 Algorithms and Data Structures

• Seeing how algorithms and data structures translate to
code helps develop your coding ability

• It give you and understanding of the basic techniques
of algorithm development, and efficiency issues

• It shows how careful attention to the application and
implementation divide gives well structured programs

• In practice, however, you wouldn’t need to write your
own code for the simple algorithms and data structures
presented here - you would use library code



 Code Re-use

• Code re-use means seeing if there is an existing
class which provides what you want (not “cut and
paste”)

• An interface with carefully chosen and well
specified methods makes code open for re-use

• Using interface types which have just the
necessary methods is one way of generalising
code, so it can be re-used in a variety of situations

• Considering how aspects of the code could be
provided as parameters rather than fixed is another
way of generalising

 Java™’s APIs

• API – “Application Programming Interface”

• The formal name for a “code library”

• Provided to give code for common operations, and
for interaction with databases, graphics, web etc

• Once you understand the basics of OOP, a lot of
programming is knowing what and how to use
from the API

• Oracle’s Java™ APIs may be supplemented by
APIs from other suppliers

• You may define and implement your own APIs