Lesson 6: Loops
by
Akos Ledeczi and Mike Fitzpatrick
The loop is a new control construct that makes it
possible to repeat a block of statements a
number of times.
We have already used loops without knowing it:
>> n = 1:5;
>> total = sum(n);
MATLAB uses loops internally both to compute
the result of the colon operator and to compute
the sum of the elements of the vector n above.
Implicit loop
Let’s compute the sum of 1 trough 5 without using
the built-in sum function!
Use a loop:
Set total to 0
Repeat for n = 1 to 5
Add n to total
Print total
Set total to 0
Set n to 1
Execute Add n to total (total equals 1)
Set total to 0
Set n to 2
Execute Add n to total (total equals 3)
Set n to 3 Repeat for n = 1 to 5
Execute Add n to total (total equals 6)
Add n to total
Set n to 4
Execute Add n to total (total equals 10)
Set n to 5
Print total
Execute Add n to total (total equals 15)
Print total
MATLAB implementation using a for-loop:
total = 0;
for n = 1:5
total = total + n;
end
fprintf('total equals %d\n',total);
total = 0;
for n = 1:5 control statement
loop total = total + n; body
end
fprintf('total equals %d\n',total);
Here is another example:
list = rand(1,5); % assigns a row vector of random numbers
for x = list
if x > 0.5
fprintf('Random number %f is large.\n',x)
else
fprintf('Random number %f is small.\n',x)
end
end
Random number 0.141890 is small.
Random number 0.421760 is small.
Random number 0.915740 is large.
Random number 0.792210 is large.
Random number 0.959490 is large.
Notice that we do not need the list variable at all:
for x = rand(1,5)
if x > 0.5
fprintf('Random number %f is large.\n',x)
else
fprintf('Random number %f is small.\n',x)
end
end
The values assigned to the loop index do not
have to be
◦ integers,
◦ regularly spaced, or
◦ assigned in increasing order,
In fact, they do not have to be scalars either:
◦ The loop index will be assigned the columns of the array
Any other control construct can be used in the
body of the for-loop
◦ if-statements
◦ other loops
◦ etc.
for-loops work well when we know the
number of necessary iterations before
entering the loop
Consider this problem:
◦ Starting from 1, how many consecutive positive
integers do we need to add together to exceed 50?
◦ The only way to solve this with a for-loop is to guess
a large enough number for the number of iterations
and then use a break statement.
◦ There is a better solution: a while-loop!
function [n total] = possum(limit)
total = 0;
n = 0;
while total <= limit
n = n + 1;
total = total + n;
end
fprintf('sum: %d count: %d\n', total, n);
function [n total] = possum(limit)
total = 0;
n = 0;
while total <= limit
n = n + 1;
total = total + n;
end
fprintf('sum: %d count: %d\n', total, n);
>> possum(50)
sum: 55 count: 10
ans =
10
function [n total] = possum(limit)
total = 0;
n = 0;
while total <= limit control statement
n = n + 1;
loop body
total = total + n;
end
fprintf('sum: %d count: %d\n', total, n);
while conditional
block
end
if conditional
block
end
Difference:
while condition is evaluated repeatedly
block is executed repeatedly as long as condition is true
Problem: given a vector, v, of scalars, create a
second vector, w, that contains only the non-
negative elements of v
Traditional solution:
w = [];
jj = 0;
for ii = 1:length(v)
if v(ii) >= 0
jj = jj + 1;
w(jj) = v(ii);
end
end
MATLAB provides a more elegant solution:
w = [];
for ii = 1:length(v)
if v(ii) >= 0
w = [w v(ii)];
end
end
The ultimate solution needs only a single line:
w = v(v >= 0);
This is an example of logical indexing.
To understand why and how this works, we
need to introduce logical arrays
