learningMatlab

Loops and controls

Logical variables

So far: bits of code get executed once with one set of inputs
sometimes want to do some things in certain conditions
and often we want to do (similar) things many times over
at the heart of this lie logical variables, branching and loops

To control this kind of behaviour, one thing we need to be able to test whether certain conditions are met. Matlab has a type of variable for this called a logical value:

logical values: true and false
these are actually functions like ones()
```
>> a = true(2,10) % 20 truths ;
```
0 and false are equivalent
non-0 and true are equivalent

test for logical equivalence with == :

>> 2 == 5       % is this true?
>> (2+3) == 5   % ... and this?
>> (2 == 5) == false   % what's this?

Logical variables, sometimes also called “booleans”, store whether a condition is “true” or “false”, 1 or 0. You can check whether one value is the same as another with the == operator, but there are also many other comparisons you can make. In Matlab-speak these are called Relational Operators.

relop	meaning
`<`	is smaller?
`<=`	is smaller or equal?
`>`	is bigger?
`>=`	is bigger or equal?
`==`	is equal?
`~=`	is not equal?

Combining logical values

to test whether A and B are true: A && B
… A or B are true: A || B
&& and || are binary operators

such as:

a = rand(1); % one random number between 0,1
% the following is true if 0.3 < a < 0.7
closeToMiddle = (a > 0.3) && (a < 0.7)

if/else

To run a part of your script only if some condition is met, we can use a logical “fork in the road”. There are several ways to do this, but probably the most important one is if/else. The following is a small code snippet that illustrates what’s going on here. The variable a gets a random value from the call to rand(). Every time you run this code, you’ll get a different random number (!). The if statement checks whether (a > 0.5) is true; if yes, then the code snippet just below it is executed; else the code in the other block is run:

% make a random number between 0, 1
a = rand(1)

% now test if this time it is > or < than 0.5
if a > 0.5
  disp('chance favours the prepared mind')
else
  disp('uhoh!')
end

Another example would be to check an input that someone has given at the command prompt. Here is a modified version of what appears in the help for the input() function:

reply = input('Do you want more? Y/N [Y]:','s');
if isempty(reply) || strcmp(reply, 'y')
  % if the reply variable is empty
  % or if the string comparision with lowercase 'y' is true
  reply = 'Y';
end

… and finally, using the trick of setting a variable to an initialization value of -1 (which has a special meaning that you can test) with if/else:

if nSubjects < 0
  disp('number of subjects ill-defined')
elseif nSubjects == 0
  disp('there were 0 subjects')
else
  disp(['there were ' num2str(nSubjects) ' subjects' ])
end

for loops

Together with branching (if/else), using for loops probably forms one of the most useful aspects of programming in Matlab (and other languages). For many tasks that you want to complete, being able to repeat the same action many times with a compact set of commands makes it a very powerful tool. Just imagine a situation where something has to be done 1000 times either by a set of key strokes and maybe copying and pasting, say in Excel, or by issuing a command like runMyScript at the Matlab prompt… :

% the syntax is quite simple:
for loopIteration = 1:10
  % on each interation, the variable
  % loopIteration takes on the next value on list
  disp('The current loop count is:')
  loopIteration
end

The key to making this work is to define clearly (but flexibly) what needs to be done over and over again. You may have to work out what it is that changes on each iteration and what needs to be repeated.

One common use for loops is to load in a whole bunch of data files from psychophysics or imaging experiments and then to group data together that way. See the Appendix_ for an exercise on exactly that.

Also keep in mind:

allocate space for variables you need to fill before loop
don’t “grow” arrays (it’s memory & time costly)
try to define the limits of your loop with variables:

  for iCounter = 1:10 % AVOID
    ... % and so on
  end

  nIterations = 10; % BETTER
  for iIteration = 1:nIterations
    ... % and so on
  end

The version using nIterations = 10 is much better, because it allows you to make other bits of your code depend on that number too (but stored in a variable). So if you change your mind later and want to iterate over 100 values instead, you can simply change things in one place.

Debugging other people’s code - find the problem.

This site is open source. Improve this page.