SECTION 5:

STRUCTURED
PROGRAMMING IN MATLAB
ENGR 112 – Introduction to Engineering Computing
2 Conditional Statements
• if statements
• if…else statements
• Logical and relational operators
• switch…case statements

K. Webb ENGR 112

 The if Statement
3

 We’ve already seen the if structure

 If X is true, do Y, if not, don’t do Y
 In either case, then proceed to do Z
 In MATLAB:
if condition
statements
end
 Statements are executed if condition is true
 Condition is a logical expression
 Either true (evaluates to 1) or false (evaluates to 0)
 Makes use of logical and relational operators
 May use a single line for a single statement:
if condition, statement, end
K. Webb ENGR 112
 Logical and Relational Operators
4

Operator Relationship or Logical Operation Example

== Equal to x == b
~= Not equal to k ~= 0
< Less than t < 12
> Greater than a > -5
<= Less than or equal to 7 <= f
>= Greater than or equal to (4+r/6) >= 2
NOT– negates the logical value of an
~ expression
~(b < 4*g)
AND – both expressions must
& or && evaluate to true for result to be true
(t > 0)&&(c == 5)
OR – either expression must
| or || evaluate to true for result to be true
(p > 1)||(m > 3)
K. Webb ENGR 112
 Short-Circuit Logical Operators
5

 Note that there are two AND and two OR operators

available in MATLAB
 AND: & or &&
 OR: | or ||

 Can always use the single operators: & and |

 The double operators are short-circuit operators
 Only evaluate the second expression if necessary –
faster
 Can only be used with scalar expressions

K. Webb ENGR 112

 The if…else Structure
6

 The if … else structure

 Perform one process if a condition
is true
 Perform another if it is false

 In MATLAB:
if condition
statements1
else
statements2
end

K. Webb ENGR 112

 The if…elseif…else Structure
7

 The if … elseif … else structure

 If a condition evaluates as false, check another condition
 May have an arbitrary number of elseif statements

 In MATLAB:
if condition1
statements1
elseif condition2
statements2
else
statements3
end
K. Webb ENGR 112
 The if…else, if…elseif…else Structures
8

 Some examples:

 Note that && and || are used here, because

expressions involve scalars
 The single logical operators, & and |, would work just
as well
K. Webb ENGR 112
 The if…elseif Structure
9

 We can have an if statement

without an else
 Similarly, an if…elseif
structure need not have an
else
 In MATLAB:
if condition1
statements1
elseif condition2
statements2
end
K. Webb ENGR 112
 The switch Structure
10

 The switch structure evaluates a single test expression

 Branching determined by the value of the test expression

switch testexpression
case value1
statements1
case value2
statements2
otherwise
statements3
end

 An alternative to an if…elseif…else structure

K. Webb ENGR 112

 The switch Structure
11

 An example – set the value of variable B to different

values depending on the value of variable A:

 otherwise serves the same purpose as else

 If the test expression does not equal any of the specified
cases, execute the commands in the otherwise block
K. Webb ENGR 112
 The switch Structure
12

 In flowchart form, there is no direct translation for the switch

structure
 We’d represent it using an if…elseif…else structure
 But, if there were, it might look something like this:

switch A
case 1
B = 2;
case 2
B = 8;
case 3
B = -5;
otherwise
B = 84;
end

K. Webb ENGR 112

 The switch Structure
13

 An alternative to an if…elseif…else structure

 Result is the same
 Code may be more readable

K. Webb ENGR 112

14 while Loops

K. Webb ENGR 112

 The while loop
15

 The while loop

 While X is true, do A
 Once X becomes false, proceed to B

 In MATLAB:
while condition
statements
end
 Statements are executed as long
as condition remains true
 Condition is a logical expression
K. Webb ENGR 112
 while Loop – Example 1
16

 Consider the following while loop example

 Repeatedly increment x by 7 as long as x is less than or equal to 30
 Value of x is displayed on each iteration, due to lack of output-
suppressing semicolon

 x values displayed: 19, 26, 33

 x gets incremented beyond 30
 All loop code is executed as long as the condition was true at the
start of the loop

K. Webb ENGR 112

 The break Statement
17

 Let’s say we don’t want x to increment beyond 30

 Add a conditional break statement to the loop

 break statement causes loop exit before executing all code

 Now, if (x+7)>30, the program will break out of the loop and
continue with the next line of code
 x values displayed: 19, 26
 For nested loops, a break statement breaks out of the current
loop level only
K. Webb ENGR 112
 while Loop – Example 1
18

 The previous example could be simplified by modifying the

while condition, and not using a break at all

 Now the result is the same as with the break statement

 x values displayed: 19, 26
 This is not always the case
 The break statement can be very useful
 May want to break based on a condition other than the loop condition
 break works with both while and for loops

K. Webb ENGR 112

 while Loop – Example 2
19

 Next, let’s revisit the while loop

examples from Section 4
 Use input.m to prompt for input
 Use display.m to return the result

K. Webb ENGR 112

 while Loop – Example 3
20

 Here, we use a while loop to

calculate the factorial value of a
specified number

K. Webb ENGR 112

 while Loop – Example 3
21

 Add error checking to ensure that x

is an integer
 One way to check if x is an integer:

K. Webb ENGR 112

 while Loop – Example 3
22

 Another possible method for

checking if x is an integer:

K. Webb ENGR 112

 Infinite Loops
23

 A loop that never terminates is an infinite loop

 Often, this unintentional
 Coding error
 Other times infinite loops are intentional
 E.g., microcontroller in a control system
 A while loop will never terminate if the while condition
is always true
 By definition, 1 is always true:

while (1)
statements repeat infinitely
end

K. Webb ENGR 112

 while (1)
24

 The while (1) syntax can be used in conjunction with a

break statement, e.g.:

 Useful for
multiple break
conditions
 Control over
break point
 Could also
modify the
while condition

K. Webb ENGR 112

25 for Loops

K. Webb ENGR 112

 The for Loop
26

 The for loop

 Loop instructions execute a specified
number of times
 In MATLAB:
for index = start:step:stop
statements
end

 Note the syntax – looks like a vector

definition
 Statements are executed once for each
element in the vector
 However, index is actually a scalar
 Increments through the vector of values

K. Webb ENGR 112

 for Loop – Example 1
27

 Next, we’ll revisit the for loop

examples from Section 4
 Loop iterates 5 times
 Value of scalar variable, x, reassigned
on each iteration

K. Webb ENGR 112

 for Loop – Example 2
28

 Here, x is defined as a vector

 Loop still iterates 5 times
 Successive values appended to the
end of x
 x grows with each iteration

K. Webb ENGR 112

 for Loop – Example 3
29

 In this case the loop counter is not

used at all within the loop
 Random number generated on each
of 10 iterations

K. Webb ENGR 112

30 Nested Loops

K. Webb ENGR 112

 Nested Loop – Example 1
31

 Recall the nested for loop example

from Section 4
 Generate a matrix 𝐶𝐶 whose entries
are the squares of the elements in 𝐵𝐵
 Nested for loop
 Outer loop steps through rows
 Counter is row index
 Inner loop steps through columns
 Counter is column index

K. Webb ENGR 112

 Nested Loop – Example 1
32

K. Webb ENGR 112

 Nested for Loop – Example 2
33

 Evaluate a function of two variables:

−𝑥𝑥 2 −𝑦𝑦 2
𝑧𝑧 = 𝑥𝑥 ⋅ 𝑒𝑒
over a range of −2 ≤ 𝑥𝑥 ≤ 2 and −2 ≤ 𝑦𝑦 ≤ 2

 A surface in three-
dimensional space
 In Section 7, we’ll learn how
to generate such a plot

K. Webb ENGR 112

 Nested for Loop – Example 2
34

−𝑥𝑥 2 −𝑦𝑦 2
𝑧𝑧 = 𝑥𝑥 ⋅ 𝑒𝑒
 Evaluate the function over a range of 𝑥𝑥 and 𝑦𝑦

 First, define x and y

vectors
 Use a nested for
loop to step through
all points in this
range of the x-y
plane
K. Webb ENGR 112
35 The MATLAB Debugger

K. Webb ENGR 112

 Debugging
36

 You’ve probably already realized that it’s not uncommon for your
code to have errors
 Computer code errors referred to as bugs
 Three main categories of errors
 Syntax errors prevent your code from running and generate a MATLAB
error message
 Runtime errors – not syntactically incorrect, but generate an error upon
execution – e.g., indexing beyond matrix dimensions
 Algorithmic errors don’t prevent your code from executing, but do
produce an unintended result
 Syntax and runtime errors are usually more easily fixed than
algorithmic errors
 Debugging – the process of identifying and fixing errors is an
important skill to develop
 MATLAB has a built-in debugger to facilitate this process

K. Webb ENGR 112

 Debugging
37

 Identifying and fixing errors is difficult because:

 Programs run seemingly instantaneously
 Incorrect output results, but can’t see the intermediate
steps that produced that output

 Basic debugging principles:

 Slow code execution down – allow for stepping through
line-by-line
 Provide visibility into the code execution – allow for
monitoring of intermediate steps and variable values

K. Webb ENGR 112

 MATLAB Debugger – Breakpoints
38

 Breakpoint – specification of a line of code at which

MATLAB should pause execution
 Set by clicking on the dash to the left of a line of code in
an m-file
 MATLAB will execute the m-file up to this line, then pause

 Clicking here sets a

breakpoint
 Indicated by red
circle

K. Webb ENGR 112

 MATLAB Debugger – Breakpoints
39

 Click Run to begin execution

 Execution halts at the
breakpoint
 Before executing that line

 Command window prompt

changes to K>>
 Can now interactively enter
commands

 Toolbar buttons change

from RUN to DEBUG
K. Webb ENGR 112
 MATLAB Debugger – Breakpoints
40

 Click Step to execute the

current line of code
 Green arrow indicator
advances to the next line
 Variable, m, defined on
previous line (line 16) is
now available in the
workspace
 Can be displayed in the
command window

K. Webb ENGR 112

 Debugger – Example
41

 Recall a previous example of an algorithm to square every

element in a matrix
 Let’s say we run our m-file and get the following result:

 Resulting matrix is transposed

 Use the debugger to figure out why
K. Webb ENGR 112
 Debugger – Example
42

 Set a breakpoint in the

innermost for loop
 Click Run, code executes
through the first iteration
of the inner for loop
 Workspace shows i=1
and j=1
 Display B(i,j) and
C(i,j) in the command
window
 Both are as expected
K. Webb ENGR 112
 Debugger – Example
43

 Click Continue, code

executes until it hits the
breakpoint again
 One more iteration of
inner for loop
 Now, i=1 and j=2
 First row, second column
 B(i,j) = 2, as
expected
 But, C(i,j) = 81
 Should be 4
K. Webb ENGR 112
 Debugger – Example
44

 We see that C(1,2) is being set to B(2,1)^2

 This leads us to an error on line 21 of the code

K. Webb ENGR 112

45 Miscellany

K. Webb ENGR 112

 Sections
46

 Define sections within an m-file

 Execute isolated blocks of code
 Starts with a double comment
 Ends at the start of the next
section
 Useful for debugging, particularly
if running the entire m-file is
time-consuming
 To run a section:
 Place cursor in section and type
Ctrl+Enter
 Click the Run Section button

K. Webb ENGR 112

 Preallocation
47

 Note the red line to the right of line 14 and the red squiggle
under x in the following for loop:

 Mouse over the line or the squiggle to see the following

warning:

 The size of x grows with each iteration of the loop

 Inefficient - slow
K. Webb ENGR 112
 Preallocation
48

 When you assign a variable, MATLAB must store it

in memory
 Amount of memory allocated for storage depends on
the size of the array
 If the variable grows it must be copied to a new, larger
block of available memory – slow
 If the ultimate size of a variable is known ahead of
time, we can preallocate memory for it
 Assign a full-sized array of all zeros
 Overwrite elements on each iteration
 Array size remains constant

K. Webb ENGR 112

 Preallocation – Example
49

 A nested for loop stepping

through an 𝑁𝑁 × 𝑁𝑁 matrix
 Here N = 100

 Time the loop with and

without preallocation
 Use tic … toc

 Preallocation speeds up
the loop up significantly
 But …

K. Webb ENGR 112

 Preallocation – Example
50

 An accurate comparison
must account for the cost
of preallocation
 Start the timer before
preallocating
 Still significantly faster,
even accounting for
preallocation
 Note that times vary from
run to run
 But …

K. Webb ENGR 112

 Preallocation – Example
51

 6 msec vs. 9 msec? So what?

 Difference is imperceptible

 Now, increase N to 5e3

 25e6 elements in A!
 A significant, and very
noticeable, difference
 Preallocation is always a
good practice

K. Webb ENGR 112