Chapter 3: Linked Lists

Objectives

Looking ahead – in this chapter, we’ll consider:

• Singly Linked Lists
• Doubly Linked Lists
• Circular Lists
• Skip Lists
• Self-Organizing Lists
• Sparse Tables
• Lists in the Standard Template Library

Data Structures and Algorithms in C++, Fourth Edition 2

 Introduction
• Arrays are useful in many applications but suffer from two
significant limitations
– The size of the array must be known at the time the code is compiled
– The elements of the array are the same distance apart in memory,
requiring potentially extensive shifting when inserting a new element
• This can be overcome by using linked lists, collections of
independent memory locations (nodes) that store data and
links to other nodes
• Moving between the nodes is accomplished by following the
links, which are the addresses of the nodes
• There are numerous ways to implement linked lists, but the
most common utilizes pointers, providing great flexibility
Data Structures and Algorithms in C++, Fourth Edition 3
 Singly Linked Lists
• If a node contains a pointer to another node, we can string
together any number of nodes, and need only a single
variable to access the sequence
• In its simplest form, each node is composed of a datum and
a link (the address) to the next node in the sequence
• This is called a singly linked list, illustrated in figure 3.1
• Notice the single variable p used to access the entire list
• Also note the last node in the list has a null pointer ( \ )

Data Structures and Algorithms in C++, Fourth Edition 4

 Singly Linked Lists (continued)

Fig. 3.1 A singly linked list

• The nodes in this list are objects created from the following
class definition

Data Structures and Algorithms in C++, Fourth Edition 5

 Singly Linked Lists (continued)
class IntSLLNode {
public:
IntSLLNode() {
next = 0;
}
IntSLLNode(int i, IntSLLNode *in = 0){
info = i; next = in;
}
int info;
IntSLLNode *next;
}
• As can be seen here and in the previous figure, a node
consists of two data members, info and next

Data Structures and Algorithms in C++, Fourth Edition 6

 Singly Linked Lists (continued)
• The info member stores the node’s information content; the
next member points to the next node in the list
• Notice that the definition refers to the class itself, because the
next pointer points to a node of the same type being
defined
• Objects that contain this type of reference are called self-
referential objects
• The definition also contains two constructors
– One sets the next pointer to 0 and leaves info member undefined
– The other initializes both members
• The remainder of figure 3.1 shows how a simple linked list can
be created using this definition
Data Structures and Algorithms in C++, Fourth Edition 7
 Singly Linked Lists (continued)

Fig. 3.1 (continued) A singly linked list

• This example illustrates a disadvantage of single-linked lists: the

longer the list, the longer the chain of next pointers that need
to be followed to a given node
• This reduces flexibility, and is prone to errors
• An alternative is to use an additional pointer to the end of the
list, as seen in figure 3-2 (pages 78-80) and figure 3-3(a)
Data Structures and Algorithms in C++, Fourth Edition 8
 Singly Linked Lists (continued)
• The code uses two classes
– IntSLLNode, which defines the nodes of the list
– IntSLList, which defines two pointers, head and tail, as well as
various member functions to manipulate the list
• An example of this list in shown in figure 3-3

Fig. 3-3 A singly linked list of integers

Data Structures and Algorithms in C++, Fourth Edition 9

 Singly Linked Lists (continued)
• While figure 3-3a shows the structure of the list, a more
common representation is shown in figure 3-3b
• Let’s consider some common operations on this type of list
• Insertion
– Inserting a node at the beginning of a list is straightforward
– First, a new node is created (figure 3-4a)
– The info member of the node is initialized (figure 3-4b)
– The next member is initialized to point to the first node in the list,
which is the current value of head (figure 3-4c)
– head is then updated to point to the new node (figure 3-4d)

Data Structures and Algorithms in C++, Fourth Edition 10

 Singly Linked Lists (continued)

Fig. 3-4 Inserting a new node at the beginning of a singly linked list

– Note that if the list is initially empty, both head and tail would be
set to point to the new node

Data Structures and Algorithms in C++, Fourth Edition 11

 Singly Linked Lists (continued)
• Insertion (continued)
– Inserting a node at the end of a list is likewise easy to accomplish as
illustrated in the next slide
– The new node is created and the info member of the node is
initialized (figures 3-5a and 3-5b)
– The next member is initialized to null, since the node is at the end of
the list (figure 3-5c)
– The next member of the current last node is set to point to the new
node (figure 3-5d)
– Since the new node is now the end of the list, the tail pointer has to
be updated to point to it (figure 3-5e)
– As before, if the list is initially empty, both head and tail would be
set to point to the new node

Data Structures and Algorithms in C++, Fourth Edition 12

 Singly Linked Lists (continued)

Fig. 3-5 Inserting a new node at the end of a singly linked list

Data Structures and Algorithms in C++, Fourth Edition 13

 Singly Linked Lists (continued)
• Deletion
– The operation of deleting a node consists of returning the value stored
in the node and releasing the memory occupied by the node
– Again, we can consider operations at the beginning and end of the list
– To delete at the beginning of the list, we first retrieve the value stored
in the first node (head → info)
– Then we can use a temporary pointer to point to the node, and set
head to point to head → next
– Finally, the former first node can be deleted, releasing its memory
– These operations are illustrated in figure 3.6(a) – (c)

Data Structures and Algorithms in C++, Fourth Edition 14

 Singly Linked Lists (continued)
• Deletion (continued)

Fig. 3-6 Deleting a node at the beginning of a singly linked list

– Two special cases exist when carrying out this deletion

– The first arises when the list is empty, in which case the caller must be
notified of the action to take
– The second occurs when a single node is in the list, requiring that
head and tail be set to null to indicate the list is now empty
Data Structures and Algorithms in C++, Fourth Edition 15
 Singly Linked Lists (continued)
• Deletion (continued)
– Deleting at the end of a list requires additional processing
– This is because the tail pointer must be backed up to the previous
node in the list
– Since this can’t be done directly, we need a temporary pointer to
traverse the list until tmp → next = tail
– This is illustrated in figures 3-7 (a) – (c) in the next slide
– Once we have located that node, we can retrieve the value contained
in tail → info, delete that node, and set tail = tmp
– This is illustrated in figures 3-7 (d) – (f) in the next slide

Data Structures and Algorithms in C++, Fourth Edition 16

 Singly Linked Lists (continued)

Fig. 3-7 Deleting a node from the end of a singly linked list

Data Structures and Algorithms in C++, Fourth Edition 17

 Singly Linked Lists (continued)
• Deletion (continued)
– The same special cases apply to deleting a node from the end of a list
as they did to deleting a node from the beginning of a list
– Now these deletion operations simply delete the physically first or last
node in the list
– What if we want to delete a specific node based on its info member?
– In that case we have to locate the specific node, then link around it by
linking the previous node to the following node
– But again, to do this we need to keep track of the previous node, and
we need to keep track of the node containing the target value
– This will require two pointers, as shown in figure 3-8 (a) – (d)

Data Structures and Algorithms in C++, Fourth Edition 18

 Singly Linked Lists (continued)

Fig. 3-8 Deleting a node from a singly linked list

– As can be seen, the two extra pointers, pred and tmp, are initialized
to the first and second nodes in the list
– They traverse the list until tmp → info matches the target value

Data Structures and Algorithms in C++, Fourth Edition 19

 Singly Linked Lists (continued)
• Deletion (continued)
– At that point, we can set pred → next = tmp → next which
“bypasses” the target node, allowing it to be deleted
– There are several cases to consider when this type of deletion is
carried out
• Removing a node from an empty list or trying to delete a value
that isn’t in the list
• Deleting the only node in the list
• Removing the first or last node from a list with at least two nodes

Data Structures and Algorithms in C++, Fourth Edition 20

 Singly Linked Lists (continued)
• Searching
– The purpose of a search is to scan a linked list to find a particular data
member
– No modification is made to the list, so this can be done easily using a
single temporary pointer
– We simply traverse the list until the info member of the node tmp
points to matches the target, or tmp → next is null
– If the latter case occurs, we have reached the end of the list and the
search fails

Data Structures and Algorithms in C++, Fourth Edition 21

 Doubly Linked Lists
• The difficulty in deleting a node from the end of a singly linked
list points out one major limitation of that structure
• We continually have to scan to the node just before the end in
order to delete correctly
• If the nature of processing requires frequent deletions of that
type, this significantly slows down operations
• To address this problem, we can redefine the node structure
and add a second pointer that points to the previous node
• Lists constructed from these nodes are called doubly linked
lists, one of which is shown in figure 3-9

Data Structures and Algorithms in C++, Fourth Edition 22

 Doubly Linked Lists (continued)

Fig. 3-9 A doubly linked list

• The code that implements and manipulates this is shown in

part in figure 3-10 (pages 91 and 92)
• The methods that manipulate these types of lists are slightly
more complicated than their singly linked counterparts
• However, the process is still straightforward as long as one
keeps track of the pointers and their relationships
• We’ll look at two: inserting and removing a node at the end of
the list
Data Structures and Algorithms in C++, Fourth Edition 23
 Doubly Linked Lists (continued)
• Insertion of a new node requires the following steps:
– First, the new node is created and the data member is initialized
– Since the node is being inserted at the end of the list, its next
member is set to null
– The prev member is set to tail to link it back to the former end of
the list
– The tail pointer is now set to point to this new node
– To complete the link, the next member of the previous node is set to
point to the new node
– These steps are illustrated in figure 3.11(a) – (f)

Data Structures and Algorithms in C++, Fourth Edition 24

 Doubly Linked Lists (continued)

Fig. 3-11 Adding a new node at the end of a doubly linked list

Data Structures and Algorithms in C++, Fourth Edition 25

 Doubly Linked Lists (continued)

Fig. 3-11 (continued) Adding a new node at the end of a doubly linked list

Data Structures and Algorithms in C++, Fourth Edition 26

 Doubly Linked Lists (continued)
• Insertion (continued)
– A special case exists if the node being inserted is the only node in the
list
– In this case there is no previous node, so both head and tail point
to the new node and in the last step, head would be set to point to
the new node
• Deletion
– Deleting a node from the end of a doubly linked list is also easy,
because there is a direct link to the previous node in the list
– This eliminates the need to traverse the list to find the previous node
– To do this, we retrieve the data member from the node, then set
tail to the node’s predecessor

Data Structures and Algorithms in C++, Fourth Edition 27

 Doubly Linked Lists (continued)
• Deletion
– The node can then be deleted, and the next pointer of the new last
node set to null
– This process is illustrated in figure 3-12(a) – (d)
– A couple of special cases need to be dealt with
• If the node being deleted is the only node in the list, head and
tail need to be set to null
• Also, if the list is empty, an attempt to delete a node should be
handled and reported to the user

Data Structures and Algorithms in C++, Fourth Edition 28

 Doubly Linked Lists (continued)

Fig. 3-12 Deleting a node from the end of a doubly linked list

Data Structures and Algorithms in C++, Fourth Edition 29

 Circular Lists
• Another useful arrangement of nodes is the circular list; in
this structure the nodes form a ring

Fig. 3-13 A circular singly linked list

• The implementation requires only one permanent pointer

(usually referred to as tail)
• Insertions at the front and the end of this type of list are
shown in figure 3-14(a) and (b)

Data Structures and Algorithms in C++, Fourth Edition 30

 Circular Lists (continued)

Fig. 3-14 Inserting nodes (a) at the front of a circular singly linked list and (b) at its end

Data Structures and Algorithms in C++, Fourth Edition 31

 Circular Lists (continued)
• Code to insert a node at the end of the list follows:
void addToTail(int el) {
if (isEmpty()) {
tail = new IntSLLNode(el);
tail->next = tail;
}
else {
tail->next =
new IntSLLNode(el,tail->next);
tail = tail->next;
}
}

Data Structures and Algorithms in C++, Fourth Edition 32

 Circular Lists (continued)
• The simplicity of this list does present a few problems
– Deleting nodes requires a loop to locate the predecessor of the tail
node, much as we saw with singly linked lists
– Operations that require processing the list in reverse are going to be
inefficient
• To deal with this, the list can be made doubly linked
• This forms two rings, one going forward through the next
pointers, and the other backwards through the prev pointers
• This is illustrated in figure 3-15

Fig. 3-15 A circular doubly linked list

Data Structures and Algorithms in C++, Fourth Edition 33

 Skip Lists
• The biggest drawback to the linked lists we’ve examined is
that they are sequential in nature
• Searching for a particular element requires we look at every
node in the list starting from head until we find the target or
reach the end of the list
• While ordering the elements on frequency of access can help,
the processing is still sequential
• To improve on this, a variation called a skip list can be used to
implement non-sequential searching of a linked list
• This structure is shown in figure 3-16

Data Structures and Algorithms in C++, Fourth Edition 34

 Skip Lists (continued)

Fig. 3-16 A skip list with (a) evenly and (b) unevenly spaced nodes of different levels; (c) the skip list with pointer nodes
clearly shown

Data Structures and Algorithms in C++, Fourth Edition 35

 Skip Lists (continued)
• As can be seen in figure 3-16(a), in addition to each node
pointing to the next node, every second node points to the
node two positions ahead, every fourth node points to the
node four positions ahead, and so forth
• In general, in a list of n nodes, for k such that 1 < k < and i
such that 1 < i < , the node at location * i points to the node at
location * (i – 1)
• This is implemented by having different numbers of pointers
in the nodes – half have one pointer, a quarter have two
pointers, and so on
• The number of pointers determines the level of the node, and
the number of levels is + 1
Data Structures and Algorithms in C++, Fourth Edition 36
 Skip Lists (continued)
• To find a target value in the list, we start with the highest level
pointers
• These are followed until we find the target, reach the end of
the list, or reach a node whose data value is larger than the
target
• In the latter two cases, we back up to the node preceding the
larger data value, go to the next level of pointers, and restart
• This process is repeated until we find the target, or reach the
end of the list or a larger value when traversing the first-level
links
• Code for the search is shown in figure 3.17 (pages 98 – 100)

Data Structures and Algorithms in C++, Fourth Edition 37

 Skip Lists (continued)
• Searching is now more efficient, at the expense of insertion
and deletion operations
• Inserting a new node, for example, requires restructuring of
the nodes that follow the new node
• In particular, the number of pointers and their values must be
changed
• To handle this, we relax the requirement on the position of
the nodes and retain the number of nodes of different levels
• This transforms the list in figure 3.16(a) to 3.16(b)
• In comparing the two, notice the number of nodes of different
levels remains the same, as required

Data Structures and Algorithms in C++, Fourth Edition 38

 Skip Lists (continued)
• The question, of course, is how this is accomplished
• Consider a list of 15 elements; there will be four levels
• Based on our earlier formulas, there will be eight nodes with
one pointer, four with two pointers, two with three pointers,
and one with four pointers
• A random number is generated to determine the level of node
to insert; the following table can be used to do this
Node
#
Level
15 4
13 – 14 3
9 – 12 2
1-8 1

Data Structures and Algorithms in C++, Fourth Edition 39

 Skip Lists (continued)
• This process is implemented by methods in the code of figure
3-17
• Another implementation question is the matter of the nodes
• A simple way to handle this would be for every node to have
the maximum number of needed pointers; this is inefficient
• The number of pointers needed is equal to the level of the
node, so we can have the node’s pointer point to an array of
pointers whose size is the level of the node
• The result of these decisions is a list whose beginning looks
like figure 3-16(c)

Data Structures and Algorithms in C++, Fourth Edition 40

 Self-Organizing Lists
• Skip lists were introduced as a technique to speed up the
searching process in lists
• Another approach is based on dynamically re-organizing the
lists as they are used
• There are several ways to accomplish this, and this section
focuses on four approaches, illustrated in figure 3-18(a) - (d):
– Move-to-front: when found, the target is moved to the front of the list
– Transpose: when the element is found, it is swapped with its
predecessor in the list
– Count: the list is ordered by frequency of access
– Ordering: the list is ordered based on the natural nature of the data

Data Structures and Algorithms in C++, Fourth Edition 41

 Self-Organizing Lists (continued)

Fig. 3-18 Accessing an element on a linked list and changes on the list depending on the self-organization technique
applied: (a) move-to-front method, (b) transpose method, (c) count method, and (d) ordering method, in particular,
alphabetical ordering, which leads to no change

Data Structures and Algorithms in C++, Fourth Edition 42

 Self-Organizing Lists (continued)
• For the first three techniques, new information is in a node
placed at the end of the list
• In the ordering technique, placement is based on maintaining
the order of the elements
• These operations are illustrated in figure 3.18 (e) and (f)

Fig. 3-18 (continued) In the case when the desired element is not in the list, (e) the first three methods add a new node with this
element at the end of the list and (f) the ordering method maintains an order on the list

Data Structures and Algorithms in C++, Fourth Edition 43

 Self-Organizing Lists (continued)
• An example of searching lists organized in this manner is
shown in figure 3-19 (page 103)
• As the example indicates, the goal of the first three
techniques is to place the most likely looked-for items near
the beginning of the list
• Ordering, on the other hand, uses properties of the data itself
to organize the list
• Counting can be looked at as ordering, but often it is simply
frequency of access and stored as separate information
• Sample runs of these techniques are shown in figure 3-20

Data Structures and Algorithms in C++, Fourth Edition 44

 Self-Organizing Lists (continued)
• Three of the techniques use both insertion at the head and
the tail of the list for comparison
• Notice there is a general improvement as the size of the file
increases
• Earlier on, the focus is on inserting new items, while later
processing is concerned with organization of existing items
• From the figures, move-to-front and count perform better
than the others
• A skip list is included for comparison; while its numbers are
quite different, the data is analyzed only in terms of
comparisons, not the effort of linking and relinking nodes

Data Structures and Algorithms in C++, Fourth Edition 45

 Self-Organizing Lists (continued)

Fig. 3-20 Measuring the efficiency of different methods using formula (number of data comparison)/(combined length) expressed in
percentages

• The conclusion to be drawn is that for smaller lists, simple

linking suffices
• As the amount of data and frequency of access increases, other
structures should be employed
Data Structures and Algorithms in C++, Fourth Edition 46
 Sparse Tables
• Tables are a data structure of choice in many applications due
to their ease of implementation, use, and access
• However, in many cases the size of the table can lead to
difficulties
• This is especially true if the table is mostly unoccupied, known
as a sparse table
• Alternative representations using linked lists can be more
useful in cases like this
• Consider the example presented in the text, where we want
to store the grades for 8000 students in 300 classes

Data Structures and Algorithms in C++, Fourth Edition 47

 Sparse Tables (continued)
• Student numbers can represent rows, and course numbers
columns
• To save space, grades can be encoded using single characters
• The three tables that support this representation are shown in
figure 3-21 (a) – (c)

Fig. 3-21 Arrays and sparse table used for storing student grades

Data Structures and Algorithms in C++, Fourth Edition 48

 Sparse Tables (continued)
• The actual grade table is in figure 3-21(d)

Fig. 3.21 (continued) Arrays and sparse table used for storing student grades

• As can be seen, this table has many open locations

Data Structures and Algorithms in C++, Fourth Edition 49

 Sparse Tables (continued)
• The table itself is 8000 (students) by 300 (classes), with one
byte per grade, totaling 2.4 million bytes
• However, if every student takes only four classes each
semester, there would only be four entries in each column,
wasting almost 99% of the total space of the table
• An alternative approach is suggested by figure 3.22, where
the data is organized into two tables
– The first, classesTaken, records each class a student takes (to a
maximum of eight), along with the student’s grade
– The second, studentsInClasses, records the students in each
class (to a maximum of 250), along with their grade

Data Structures and Algorithms in C++, Fourth Edition 50

 Sparse Tables (continued)

Fig. 3.22 Two-dimensional arrays for storing student grades

Data Structures and Algorithms in C++, Fourth Edition 51

 Sparse Tables (continued)
• The use of two tables speeds processing for various lists
• With these tables, and the assumption that an integer requires
2 bytes of storage, 417,000 bytes would be needed
• While considerably less than the original single table, it is still
wasteful, and inflexible if conditions change
• A more useful and open-ended solution utilizes two arrays of
linked lists, as shown in figure 3.23
• One array has pointers to a list of students taking a given class,
the other pointers to a list of classes taken by a student
• Each node stores the student number, class number, grade, a
pointer to the next class for the student, and a pointer to the
next student for the class
Data Structures and Algorithms in C++, Fourth Edition 52
 Sparse Tables (continued)
• If a pointer occupies two bytes, each integer two bytes, and a
character one byte, then each node is nine bytes in size
• For 8000 students and an average of four classes, this list
occupies 288,000 bytes, which is roughly 10% of the original
table and 70% of the two table variation
• Moreover, there is no wasted space, and the lists can be easily
extended

Data Structures and Algorithms in C++, Fourth Edition 53

 Lists in the Standard Template Library
• The STL implements lists as generic doubly linked lists with
head and tail pointers, as shown in figure 3.9
• The class is available in a program through the directive
#include <list>, and a list of the methods in the list
container is shown in figure 3.24 (pages 110 – 112)
• An example of their use is shown in the program in figure 3.25
(pages 112 and 113)

Data Structures and Algorithms in C++, Fourth Edition 54

 Concluding Remarks
• The motivation for introducing linked lists was to allow
dynamic allocation of memory, using only what was needed
• Insertion and deletion of data was also easily effected
• This does not eliminate the need for arrays; random access to
elements cannot be achieved with lists
• However for specific elements, or for algorithms that focus on
changing the structure of the data, lists are preferable
• Arrays also have less overhead in space utilization; all that is
needed is the space for data
• Lists need to allocate pointers, which can add substantial
space overhead depending on the application

Data Structures and Algorithms in C++, Fourth Edition 55