SCRIPTS, FLOW CONTROL, AND

DATA STRUCTURES
Programming with Matlab
CALENDAR

Unit What
1 Introduction to MATLAB
2 Flow control & Data Structures
3 Plotting
4 Strings & File IO
5 Complex Numbers
6 Combinatorics
7 Linear Algebra
8 Statistics & Data Analysis
9 Polynomial Approximation,
Curve Fitting
10 Root Finding, Numerical
Differentiation
11 Numerical Integration
12 MATLAB GUI

2
SCRIPTS

• A script is simply a file containing a series of MATLAB commands.

• Until now, we have been using MATLAB terminal to enter commands.
• For larger, more complicated tasks, it is better to write all commands in a script file.
• Upon running the script, all the commands within it will be performed sequentially
just like any C/C++ program.
• Any MATLAB command (i.e., anything that can run in the terminal) can be placed
in a script; no declarations and no begin/end delimiters are necessary.

• Scripts need to be saved to a <file>.m file before they can be run.

• You may wish to setup folders to keep scripts and functions organized.
• In order to run a script, the current folder in MATLAB must be changed to the folder
that contains the script (Command cd).
• Alternatively, this can be avoided by adding the folder’s path to MATLAB by using
Environment  Set Path (MATLAB R2012b).

3
ALGORITHM

• Outline of the necessary steps

to implement a simple algorithm, the steps would be:

1. Where does the input come from?

2. Needed formulas
3. Where does the output go?

• Get input(s)
• Calculate the result(s)
• Display the output(s)

4
SCRIPTS

• Several advantages to using scripts over the command line include:

• User can automate tasks.
• Uses smart indentation to help read and write code more comfortably.
• Allows the use of breakpoints for debugging.
• Allows the use of the MATLAB profiler and tic/toc functions to analyze
program performance.
• User can run individual pieces of code by highlighting a piece and pressing F9
(in addition to breakpoints, this makes debugging very convenient and easy).

• To remove previous variables from the workspace, you may wish to place the
command clear all at the beginning of your script.
• The close all command may also be placed at the beginning if closing all
figures is desired prior to running the script.
• It is good practice to use ; at the end of commands when writing scripts to inhibt
unwanted output
• When a script is running the >> symbol at the terminal will disappear until it
finishes, preventing any commands from being input.
• To terminate the script before conclusion and regain control, press Ctrl+C.

5
ELLIPSIS

• Word wrap for comments is automatically enabled in scripts so that each line fits in
the viewable window to prevent the need for horizontal scrolling.
• Making use of ellipsis (...) can help make scripts more readable when there are
long statements of code:

A = [1 2 3 4 5 6; ...
6 5 4 3 2 1];

• This can also be used to make long character strings by concatenating two shorter
strings together:

myString = ['Accelerating the pace of ' ...

'engineering and science'];

6
CONDITIONAL STATEMENTS

• Scripts can be made more effective when you are able to control which blocks of
code are executed.
• MATLAB is capable of flow control just like any other language.
• The most basic flow control element is the if statement.
a = 6;
if a > 10
b = 0;
elseif a < 3
b = 1;
else
b = 2;
end

• The statements are evaluated only when the logical expressions are true.
• The elseif and else are optional.

7
FOR LOOPS

• A for loop uses a counter to repeatedly execute the block of code within it.
• It follows the syntax: for index=values, statements, end.
• The values portion of the syntax can take several forms, one of which is:

A = []; ii = 2 A = 2
for ii = 2:2:8 ii = 4 A = 2 4
A = [A ii] ii = 6 A = 2 4 6
end ii = 8 A = 2 4 6 8

• In this case, the index ii is incremented by an optional step size until it reaches 8.
• A is initialized as an empty matrix and its size is constantly updated as we add to it
during each iteration of the loop.
• Note that the name of the index variable is ii rather than the traditional i,
commonly used in C/C++. Recall that i is reserved for the complex unit.
• Can create character indices as well, e.g. ii = 'a':'z'.
• The value of the index should never be modified by the statements inside the loop.

8
FOR LOOPS

• values can also be defined without the use of the colon operator:

A = zeros(1,3); %pre-allocate space iteration 1:

k = 1; %count loop iterations jj = 1
valueMatrix = [1 2 3; 5
5 6 7];
A = 6 0 0
for jj = valueMatrix
fprintf('iteration %d:\n', k) iteration 2:
A(k) = jj(1) + jj(2); jj = 2
jj, A %display variables on terminal 6
k = k + 1;
A = 6 8 0
end

• Here, the loop index is a column vector obtained from iteration 3:

subsequent columns of the matrix valueMatrix on jj = 3
each iteration. 7
• Even if column vector indices are not desired, this
method allows the use of irregular loop indices. A = 6 8 10

9
PRE-ALLOCATING SPACE TO VARIABLES

• Note that in the previous example, we pre-allocated space to matrix A rather than
letting MATLAB resize it on every iteration. Use functions like length and size
along with the colon operator to determine loop indices after pre-allocating space.

A B
tic tic
for ii = 1:2000 A = zeros(2000, 2000);
for jj = 1:2000 for ii = 1:size(A,1)
A(ii,jj) = ii + jj; for jj = 1:size(A,2)
end A(ii,jj) = ii + jj;
end end
toc end
toc

Elapsed time is 12.357 seconds Elapsed time is 2.261 seconds

• Which of the two solutions will compute faster?

10
PRE-ALLOCATING SPACE TO VARIABLES

• Note that in the previous example, we pre-allocated space to matrix A rather than
letting MATLAB resize it on every iteration (computationally expensive).
• For very large loops, it is good practice to always pre-allocate space:
A B
tic tic
for ii = 1:2000 A = zeros(2000, 2000);
for jj = 1:2000 for ii = 1:size(A,1)
A(ii,jj) = ii + jj; for jj = 1:size(A,2)
end A(ii,jj) = ii + jj;
end end
toc end
toc

Elapsed time is 12.357 seconds Elapsed time is 2.261 seconds

• Use functions like length and size along with the colon operator to determine
loop indices after pre-allocating space.

11
WHILE LOOPS

• A while loop repeats for as long as the condition remains true.

• This can be very useful for things such as numerical approximation, where we
aren’t sure of the exact number of iterations.
• E.g.: we want to approximate π using Newton’s method and we want the estimate
to be within 10-32 of the actual value:

x = 2;
k = 0;
error = inf; >> x
errorThreshold = 1e-32; x =
while error > errorThreshold 3.141592653589793
x = x - sin(x)/cos(x);
error = abs(x - pi); >> k
k = k + 1; k =
end 6

• We see that it takes 6 iterations to achieve our desired precision using an initial
estimate of x = 2.

12
BREAK

• The command break terminates the execution of for or while loops.

• Control passes to the statement that follows the end of that loop.
x = 2;
k = 0; >> x
error = inf; x =
errorThreshold = 1e-32; 3.140943912317635
while error > errorThreshold
if k > 3 >> error
break error =
end 6.4874e-04
x = x - sin(x)/cos(x);
error = abs(x - pi); >> k
k = k + 1; k =
end 4

• Can be useful in iterative methods where convergence is not guaranteed.

13
SWITCH STATEMENTS

• In a switch statement, the input is matched against several different values and a
block of code is executed for the case where the match is correct:

x = 't';
switch x
case 2
disp('x is 2')
case -1
disp('x is -1')
case 't'
disp('x is the character t')
otherwise
disp('x is not 2, -1, or t')
end

x is the character t

14
TRY/CATCH STATEMENTS

• Attempts to evaluate an expression; if an error occurs, the statement inside the

catch block is executed.
• After the error is “caught”, the script continues after the end of the try statement.
a = 1:5; c =
b = 1:3; 2
for ii = 1:length(a)
try c =
c(ii) = a(ii) + b(ii) 2 4
catch
disp('b is too small') c =
end 2 4 6
end
b is too small
b is too small

• In this example, b only has 3 elements so an error occurs when the script
attempts to access the 4th and 5th element.
• By using try/catch, MATLAB no longer terminates the script at this error.

15
SCRIPTS AND FUNCTIONS

There are two options in order to describe a sequence of MATLAB

commands on a file:
• Scripts do not accept input arguments or return output arguments. They
operate on data in the workspace. When you invoke a script, MATLAB
simply executes the commands found in the file.

• Functions can accept input arguments and return output arguments.

Internal variables are strictly local to the function. When you call a
function, MATLAB only executes the portion of the function file related to
the function and eventual sub-functions

16
FUNCTIONS

• Functions are useful for automating frequently used tasks and subroutines.
 e.g we can write a script to calculate the area of a triangle:

b = 5;
h = 3; a =
a = 0.5*(b.* h) 7.5000
• To perform this calculation again for a different triangle, we would have to modify
the values of b and h in the script and re-run it. Rather than doing this, we can
simply convert this task into a function [1].
function a = TriArea(b, h)
a = 0.5*(b.* h);
end

• The calculation can now be performed repeatedly without any modifications:

a1 = TriArea(1,5) a1 = 2.5000
a2 = TriArea(2,10) a2 = 10

17
 FUNCTION FILES

• Functions must be defined in an independent file, that must be saved in the work folder

• MATLAB files can contain code for more than one function.
• The first function in the file (the main function) is visible to functions in other files, or
you can call it from the command line.
• Additional functions within the file are called local functions. Local functions are only
visible to other functions in the same file. They are equivalent to subroutines in other
programming languages, and are sometimes called subfunctions.

• Local functions can occur in any order, as long as the main function appears first. Each
function begins with its own function definition line.

18
FUNCTIONS

• Functions have their own independent workspaces and everything that occurs
inside is hidden from the main workspace.
• From the previous example, if my main workspace has a variable b already
defined, calling Triarea will not interfere with this variable even though we use a
variable also called b inside of it.

>> b = 99; >> b

>> a3 = TriArea(2,4); b =
99
• Finally, any lines after the function name which begin with % will serve as the help
text when the function is called using the help command.

function a = TriArea(b, h)
% This function calculates
% the area of a triangle.
% INPUTS: b - base, h - height
% OUTPUTS: a – area
a = 0.5*(b.* h);
end
>> help TriArea
This function calculates
the area of a triangle.
INPUTS: b - base, h - height
OUTPUTS: a - area

19
 FUNCTIONS

Solution A: Running a simple Script

b = 5;
h = 3; a =
a = 0.5*(b.* h) 7.5000

Solution B: Calling a Function

function a = TriArea(b, h)
a = 0.5*(b.* h);
end

a1 = Triarea(5,3) a1 = 7.5000

Which of the two solutions is faster ?

20
 FUNCTIONS

Solution A: Running a simple Script

b = 5;
h = 3; a =
a = 0.5*(b.* h) 7.5000

Elapsed time is 0.000288 seconds.

Solution B: Calling a Function

function a = TriArea(b, h)
a = 0.5*(b.* h);
end

a1 = TriArea(5,3) a1 = 7.5000

Elapsed time is 0.0003390 seconds.

The first solution is faster as it does not involve seeking the file system for the function and
calling it (control overhead). Functions are useful for code readability and efficiency, but do
not speed up computation; on the contrary, they involve a little overhead

21
FUNCTION HANDLES

• Function handles provide a way to create anonymous functions, i.e., one line
expression functions that do not have to be defined in .m files.
• A function handle is a MATLAB data type and can therefore be passed into other
functions to be called indirectly.

>> f = @(x) exp(-2*x);

>> x = 0:0.1:2;
>> plot(x, f(x))

• We have assigned the function e-2x

to function handle f and evaluated
it at each point in vector x.
• To integrate e-2x from 0.2 to 0.8, we
can use f as an input argument to the
MATLAB function integral:

>> integral(f, 0.2, 0.8)

ans =
0.2342

22
CELL ARRAYS

• In a matrix, each element contains a single value and all elements in the matrix
must be of a single type, e.g. double, char, logical, etc.

• A cell array is a data structure with indexed containers (like the elements of a
matrix) called cells.

• Each cell can store any type of data and can vary in size.

• Cell arrays can be created using the same syntax as a matrix, except that curly
brackets {} are used instead of square brackets [].

>> A = {[1 2; 3 4], 27; 'Alice', {1, 2, 3}}

A =
[2x2 double] [ 27]
'Alice' {1x3 cell}

• In this example, A is a 2-by-2 cell array containing different types of arrays with
different sizes.

23
CELL ARRAYS

•To retrieve the contents of a cell, use indexing with curly brackets.

>> A = {[1 2; 3 4], 27; 'Alice', {1, 2, 3}}

>> B = A{1,1}
B =
1 2
3 4

• To retrieve a cell itself, use indexing with round brackets.

>> C = A(1,:)
C =
[2x2 double] [27]

• Here, C is a 1-by-2 cell array.

24
 STRUCTURES

• Similarly to cell arrays, structure are a method of storing heterogeneous data.

• A structure is composed of data containers called field, which are accessed by
strings.
• e.g. we want to store student assignment grades in a structure:

student.name = 'John Doe';

student.id = 'jdo2@ankara.edu.tr';
student.number = 301073268;
student.grade = [100, 75, 73; ...
95, 91, 85.5; ...
100, 98, 72];

student =

name: 'John Doe'

id: 'jd2@sfu.ca'
number: 301073268
grade: [3x3 double]
Source www.mathworks.com

25
STRUCTURES

• Records for additional students can be added by including an index after the name
of the structure:

student(2).name = 'Ann Lane';

student(2).id = 'aln4@ankara.edu.tr';
student(2).number = 301078853;
student(2).grade = [95, 95, 100; 100, 82, 85; 90, 97, 100];

student =

1x2 struct array

with fields:

name
id
number
grade

26
CELL ARRAYS VS. STRUCTURES

• Choosing between cell arrays and structures is mostly a matter of personal

preference and code clarity.
• The containers of a cell array are referred to by their index, whereas the containers
of a structure are referred to by their names.

studentData(1,:) = {'John Doe', 'jdo2@sfu.ca', 301073268, ...

[100, 75, 73; 95, 91, 85.5; 100, 98, 72]};
studentData(2,:) = {'Ann Lane', 'aln4@sfu.ca', 301078853, ...
[95, 100, 90; 95, 82, 97; 100, 85, 100]};
studentData =

'John Doe' 'jdo2@sfu.ca' [301073268] [3x3 double]

'Ann Lane' 'aln4@sfu.ca' [301078853] [3x3 double]

• Here, we accomplish what we did previously using a cell array.

• Rather than fetching the 2nd student’s no. by student(2).number, we would do
studentData{2,3}, since column 3 is where student numbers are stored.
• Since this is not as intuitive, we may conclude that a structure was a better choice.

27
MULTIDIMENSIONAL ARRAYS

• A normal 2D data structure, such as a matrix or cell array, has rows as the 1 st
dimension and columns as the 2nd.
• We may add a 3rd dimension, sometimes called a page, by adding another index.

[2]

• Just as we transitioned from a 1D vector/array to a 2D matrix, the same rules still

apply for 3D matrices and data structures.
• Additional indices can be added (dim1, dim2, dim3, ..., dimN) for
N-dimensional matrices.

28
 MULTIDIMENSIONAL ARRAYS

• There are various ways of creating multidimensional arrays:

• One way is to first create a 2D array and then extend it with indexing:

>> A = [1 2; 3 4];
>> A(:,:,2) = [9 8; 7 6]

A(:,:,1) = A(:,:,2) =
1 2 9 8
3 4 7 6

• If we assign a scalar to a new page, MATLAB fills the entire dimension with that scalar:

>> A(:,:,3) = 5

A(:,:,1) = A(:,:,2) = A(:,:,3) =

1 2 9 8 5 5
3 4 7 6 5 5

29
MULTIDIMENSIONAL ARRAYS

• They can also be generated using MATLAB functions like randn and zeros, e.g.,
to create a 4x5x2 array of zeros: A = zeros(4,5,2).

• The repmat function can be used to replicate a matrix across multiple dimensions.
• For example, to copy a matrix A across 4 pages in the 3rd dimension:

>> A = [1 2; 3 4];
>> B = repmat(A, [1 1 4])

B(:,:,1) = B(:,:,2) =
1 2 1 2
3 4 3 4

B(:,:,3) = B(:,:,4) =
1 2 1 2
3 4 3 4

30
 MULTIDIMENSIONAL ARRAYS

• Sets of matrices can be concatenated along any dimension using the cat function
C = cat(dim, A, B)

• Examples of matrix concatenation on different dimensions:

• A = [ 1 2 ; 3 4 ] B = [ 5 6; 7 8]

31
 MULTIDIMENSIONAL ARRAYS

• Sets of matrices can be concatenated along any dimension using the cat function
C = cat(dim, A, B)

>> A1 = [0 2; 4 6];

>> A2 = [1 3; 5 9];
>> B = cat(3, A1, A2)
B(:,:,1) = B(:,:,2) =
0 2 1 3
4 6 5 9

>> C = cat(4, B, B)

C(:,:,1,1) = C(:,:,2,1) =
0 2 1 3
4 6 5 9
C(:,:,1,2) = C(:,:,2,2) =
0 2 1 3
4 6 5 9

32