[MUSIC] Sometimes, during the iteration
of a for-loop or a while loop. After some of the iterations are done,
the rest of them should just be skipped. Let's look at an example of this situation
and some code that takes care of it. The code is in the file named
BreakExamples over here. And it looks a little different
from files that you've seen before. First of all, there's no function
keyword in the file, so it's a script. And also, this script, unlike the other
scripts we've seen in this course, is divided into sections. Here, let me have a little more room and
I think I'll just clear the screen here. There. This is a section, this is a section,
this is a section and so on. A section is a set of lines that begins
with two successive percent signs. The section ends when another section
begins, or at the end of the file, like this one. You remember that the percent
sign signifies that it and the text that follows it on the same
line are part of a comment, which is there only for humans to read. MATLAB colors a comment green. When there are two percent signs in a row,
it means that a new section is beginning. MATLAB indicates that there's something
special about that with a bold font. This font and
this font is bold compared to that one. The significance of sections is that
each section can be run by itself, as if it were the only
thing in the script. You do that using special
buttons on the editor. They're up here,
we'll look at them in just a minute. But first, let's look at
the first section of this script. Not this one, actually the second section,
which has a while loop in it. It sets all the values in the vector
named readings, this vector here equal to 0 until it's reached
the first value with exceeds 100. The condition here in the while loop
makes this happen with two parts. The first part here before the ampersand,
ampersand, allows the loop to continue only while the index is not greater
than the length of readings. And the second part allows it to continue
only while the readings are less than or equal to 100. You may remember that a major
difference between a script and a function is that a script has no
arguments, and it has no local variables. It uses the variables in
the command window workspace. Since the script here reads
the variable readings, we need to give readings a value. Let's give it a row vector
of 20 random integers, and we'll reset the random number generator
first, so we can regenerate the same set of pseudo-random numbers for
the next two examples, and so that you can easily repeat this
exercise with the same results. I've generated numbers, you know,
ranging from 1 to 105 here, want to get at some numbers above 100. And I see I've got some here. First one appears to be on one,
two, three, four, five, six, seven, eight, at element nine. Now let's run example one. I click anywhere inside this section. And when I click in that section,
it's highlighted in yellow. Then I come up here to the ribbon, and I click the small green arrow with the
tiny yellow page of text behind it here. It says, Run Section. Well, it didn't seem like
anything happened but let's take a look at readings. Ah-hah. We can see that the one, two,
three, four, five, six, seven, first eight elements are now zero. And nothing has happened beyond there,
because, over here the while loop got to element
nine, realized it wasn't less than or equal to 100, so
it skipped the rest of the iterations. It did the first eight iterations, skipped all the rest of them,
which is what we wanted it to do. Well, this works fine. But, the condition of the while-loop
seems a bit cumbersome, all this here. Or, at least let's call it cumbersome to
give us an excuse to show you another way to do this. Example two provides another way. Let me highlight it. First of all, it uses a for loop instead
of a while loop, which you can see here. Its control statement easily enforces
the requirement that the index stay within the range of indices of
the vector readings. But there is no possibility for the control statement of any for-loop
to check any condition whatever, the way there is with a control
statement of a while loop up here. So it seems at this point,
that the loop will scan all the elements. The body of this for-loop contains
just one if else statement. That statement checks the value of element ii readings here to see if
it's greater than zero. If it is,
it executes this break statement. Otherwise, it sets
the element equal to zero. Word break here is highlighted in
blue because it's a MATLAB key word. And this same key word is used in
the same way in many other languages. It's not the same thing as a break point,
these things we've set when we click up here and make a red circle to
tell the debugger to stop on a given line. Now this break statement is an executable
statement that causes the loop to end immediately. And it works for
both while loops and for-loops. And it's a more powerful mechanism for
ending a loop than a while condition because, in addition to skipping
the subsequent interations, it can skip the rest of the statements
in the current iteration. Here for example, there could be more
statements after this if else statement. And every one of them would be skipped
if this break statement were executed. In short, the loop quits as soon as that
break statement is executed regardless of where it occurs in the loop body. Let's try this example. First, we give our vector
readings the same pseudo random number sequence as before. See if I can find that, there. And now we click anywhere inside the next
section of code, this one, example two. Last time we did example one. And we hit the run section button again. Now, let's see if it worked. We have to look at readings. And it looks like it worked again. All these values before 101, which were originally less than 100, are now set to 0, and
the rest have been untouched. Let's go over one more time how it worked. We enter the for loop, and it appears that
we're going to process each element of the vector because of the phrase
ii equals 1:length of readings. That doesn't happen. As long as this if statement
finds that the value of element ii of readings
is not greater than 100, then the else clause dutifully
sets that element to 0. When the if else statement
finds that the value is greater than 100,
it enters the if clause and hits the break right here which
ends the loop immediately. By the way the break statement
can appear only inside a loop. In fact Matlab will stop the program and print an error message if it encounters
a break anywhere that's not inside a loop. Okay, both these examples work,
but we've got another example. It also uses the break statement
as you can see here, but it takes advantage of Matlab's answer to an age
old question in programming languages. When the loop ends,
what happens to the for loop index? Well in most languages and Matlab is
one of them, it's still there and it keeps the last value it had
while the loop was executing. And that's true whether the for
loop ended normally or ended with a break and
whether it was set by the for loop control statement or
by a statement inside the body, et cetera. Occasionally we can take advantage of
the fact that the loop index keeps its last value. For example, suppose the problem
is simply to find the first index with a value greater than 100,
without setting smaller values equal to 0. Example 3 was designed to do just that. It starts out with the apparent intention
of scanning through all the elements, of the readings,
with ii=1:length of readings. Its body is an if statement though. And it uses a break statement to exit
the loop at the first element that's greater then 100. After the loops comes this fprintf
statement which tells us the index of the first reading above 100. Looks pretty good, let's try it. We'll generate the same values once again. You recognize them by now. And we click in the last section, and then click the Run Section
button to run the code. And over here we see it says,
first reading above 100 is at index 9, and that's correct. This seems to work well. But there's a bug lurking in this code. Can you see it? It will give the wrong
answer in many cases. Let's try one of those cases. We just need to give it values that
are all less than or equal to 100. There, now let's run it again. Hm, it seems to think that the last
element is greater than 100. But, it's not. Well, let me see if I can
figure out the problem. Gosh, I'm kind of tired, though. Hey, I know what. I'll let you fix it. It'll be good exercise for
you to exterminate this bug. And, since you're going
to take care of that, we're done with our little script and
its sections. I got to show you one feature of scripts
with sections that is quite cool. Let's go up here to the ribbon and
click on the tab labeled Publish here. There are lots of things you can do here,
but I'm going to show you just one. Let's click this button labeled Publish. And what shows up is a web
page version of your script with fancy type setting and
layouts and links and so on. Let's compare the two. Here we have the top label,
might call it a title. It's been put in red, nice, large type. The version of Matlab you have
may vary this a little bit. This is the latest version
as I'm speaking here. And then there are these sections here. Each one of those is put inside
this thing called Contents, and if we click on these we
see that they are links. Like, there's Example 1. If we click there, we're on Example 1. You can see the code. Here's the comment I've got in Example 2. Skipping is accomplished
with a break statement. That's go to Example 3 down here. Shows that the for
loop index looks very professional and even runs each one of these. There was no output with the first two. But this one has output. First reading above 100 is at index 20. Hm, well,
that reminds me that bug is still there. You haven't fixed it yet. Maybe we don't want to show this. But if we did,
you can link it into other web pages. You can email it. You can show it to your friends. I think that's quite cool. And by the way, this webpage is stored in
a little folder inside the current folder. That's generated by publish. If you go in here,
you can see BreakExamples, which was the name of this m file, .html. Let's go back up to where we were before. Okay now, what were we talking about? Oh yes, you were going to get rid
of that embarrassing bug, you know, that we discovered when we
ran this third example. And speaking of bugs, there's an extremely
common bug that shows up when break statements are used in nested loops. It's a big problem for novice programmers. Not that, you know, being a novice
programmer is like a bad thing. Me and my favorite people over
the years have been novice programmers. I used to be a novice programmer myself. The fact remains that novice programmers
have trouble with this problem. The problem is that the break statement
breaks out of only the innermost loop. Meaning that it will cause the loop
it appears in to terminate, but the outer loop will continue. That is often not what
the new programmer expects. We're going to look at an example that
works with a two-dimensional array. So let's generate one. There is a nice array. We want to look at each element in
row measure order, that's first row, then second row and so on, and
set the elements equal to zero until we find the first value that's
greater than 90, and we want to leave that element and all
the other elements after it as they are. So in this case,
we want to put all zeros on this first row because there's nothing
greater than 90 there. Then we move the second row,
and we put zero here and here, but when we get to the 97,
we stop, we don't put a zero there, and we se, leave everything else as it is,
and we're done. This is a job for for-loops, and since we want to be able to stop before we
get to the end, but we can't know where we want to stop until we get there,
it's also a job for the break statement. Here's our first try in
this file called bad break. [SOUND] As you can see, there's a for loop here whose index ii goes
from 1 to the size of A,1. Size of A,1 means the first dimension
of A, which is the number of rows. Then there's an enter loop whose index
jj goes from one to the size of A,2. Size of A,2 returns the second dimension
of A, which is the number of columns. So ii is set equal to one,
and we go in here, and jj goes from one to eight because
there are eight columns in A. And it checks the values of
the elements ii, jj against 90, and if they are less than or equal to 90,
it sets them to zero, else it breaks. Well, that looks pretty good. Let's try it. We can run it by typing its name. We can also run it by
clicking on run up here. That works well for
scripts, so let's do that. So bad break ran and
let's take a look at A. Something's wrong. We should've set the first row
equal to zeros, we did that. Then we should've set these two
equal to zero, we did that, and stopped with the 97, and
it looks like we're stopped. But then, we seem to start up again
on the third row, two more zeroes. And wait a minute,
this 63 here got set to zero, too. This is exactly the problem
we were talking about. The break statement stopped the inner
loop, but not the outer loop. Well, this example is just complicated
enough to warrant the debugger. First, I'm going to clear the command
window because I want to print some things in there, and then I'm going to set
A back to the value it had before. [SOUND]
There. Okay, I'm going to set a break point,
but before I do, I'm going to remove the semicolon here, after this
zero because I want A to print out. And I want it to print out each
time through the inner loop, and that will make it do it. Now I'm ready to set a break point, I'm going to put it right
here at the beginning. Woo, it's gray,
that means I have to save this. It means that the file that I'm
breaking is not the same file that I'm, that I'm looking at here. Okay, we're ready. Let's run it. I'm going to click run. And we see that bad break is run and
K means we are in the break system, and so this little green arrow,
which points to the next statement, is going to be executed. We haven't done anything yet. We are about to enter the outer loop,
so I'm going to take one step. When I do this, you see that little hint
there that that shows you, depending on your operating system, a shortcut you
can use for this on the keyboard. I like to use the button,
so I'm going to do that. So we go in here, and
ii is set equal to one. If you hover over the ii you
can see its value is one. Now we're about to enter the inner loop. Let's do that. Now we see that ii is still one,
jj is one, and we're checking the value of A,
ii, jj, which is A one one. It's 81.
If it's less than or equal to 90, we'll set it equal to zero. And of course it is, so
we're going to enter this if branch and set this equal to zero,
so let's carry that out. And when we do,
you'll see the whole array A printed out. That's what happens when you
assign any value to an array and leave the semicolon off
the assignment statement. Not just the element that you assigned,
but the whole array prints out, and that's kind of nice for us here. We can see what the array looks like and we can see how it was
changed from the last time. Just this zero showed up here. Okay, we're ready to complete
that iteration of the inner loop. And now,
here we are at the second iteration. The outer loop still has ii equal to one,
the inner loop now has it at two, and we're about to check the element one two,
which is also less than 90. And you can see it's set to zero. And we can continue in this way,
going all the way through, the first row. Every single thing on that
row gets set to zero. Now we finish that row, and that puts us at the end of the inner loop,
and we're down here at the end statement
of the outer loop, and you know, the end statement will cause the ii to
take on its next value, which is two. So lets go forward one step. Okay, we're now back at the top, and
we're about to enter the inner loop again. This time ii is equal to two. Be a little more careful here. So now we've entered that loop,
ii is equal to two, jj is equal to one. So we're going to be looking at this
element right here, which is equal to 90. It is less than or
equal to 90 because it's equal to 90, so we'll go into this branch again. There we did, and we'll carry it out,
by setting that element equal to zero. Let's do that. And it got set equal to zero,
as you can see here. So we're at the end of that iteration. Let's take another step. Ii's still equal to two,
we're still on the second row. Jj is now equal to two, so
we're looking at this element, it's 28, it'll be less than 90. So we enter the if branch, we set it equal
to zero, you saw it become zero there. Now lets take one more step. We've entered the third
iteration of this inner loop. Ii is equal to two, jj is equal to three,
so we're about to look at this element. This element is not less than or
equal to 90, it's 97, so we'll go to the else branch. There we are. We're about to execute
the break statement. In the break statement, we were
expecting to exit the whole thing but it's only going to exit the inner loop. So let's watch that happen. There.
We jump down. We're out of the inner loop. We've jumped down to the next iteration
of the outer loop, ii is now equal to 3. So we're down here on the third row. Now we see the problem. We needed to break out of the outer loop at the same time we broke
out of the inner loop. We needed to break out of this loop
when we broke out of this loop, with this break statement. Well, now we know what went wrong,
so let's quit our debugging session. We can do that by clicking the great big
red square here labeled Quit Debugging. There, bugging session is over,
and one indication that it is, is there's no green arrow here, and another one is that we don't have
the K in front of our prompt anymore. We've established that we need
to break out of the outer loop as well as the inner loop when
we find what we're looking for. So how do we do that? Well we've got code here
that does it in GoodBreak.m. Let's take a look at that. Well first,
we see a new variable here called found. We set it equal to false before
we even enter the outer loop. Then, inside the inner loop down
here is where we have the break. We set found equal to true
before we execute the break. And then in the outer loop, after we're done with the inner loop,
we check found to see if it's true. There are two versions of the conditional
for checking the value of found. The first version is found
double equals true, like this. That certainly checks to see
if found is equal to true. And the second version is
the one we've got here. Either case works because in both cases,
the condition will be true if and only if found is equal to true. If found is true,
then we break out of the loop. We've already broken out of the inner loop
right after we set found equal to true, and so we break out of the outer loop and
we're done. The variable found which shows up in
three places typically is called a flag. A flag usually has only a few values,
and most often just two values, like this, true and false the most common. And it serves as a signal that
affects the behavior of the program. Here, the behavior is to
exit the outer loop or not. And the flag is the standard solution
to the problem of exiting nested loops. There is another way, but
it can only be used when a, the nested loop is inside a function and b, it's appropriate to exit the function
when you exit the nested loops. You just replace this break in
the inner loop with a return. And you can forget this part. And this part. And this part. In other words you can forget all
the breaks and all the flags. Return leaves the function for good, regardless of how many
nested loops there are. If you have work to do in the function
after the nested loops are done, then it's not appropriate to exit the function
when you exit the nested loops and it typically is inappropriate, so
you should remember how to use a flag. Well, let's get our nested loops
back the way they were by doing a few of these undos. There. And with that,
we're done with the break statement, but there is a related statement
called the continue statement. It's similar to the break but instead of using the break keyword,
you use the continue key word. And instead of quitting the loop, it skips
what's left of the current iteration of the loop and
moves to the next iteration. In a for-loop the loop index takes on its
next value and next iteration begins. In a while loop,
it just goes to the beginning and checks the loop condition again. What probably the biggest difference
between the continue statement and the bracket statement, is that the
continue statement is just not used very much, and
the break statement is used a lot. So we're not going to cover the continue
statement in this course, but if you're interested in it,
check it out in the textbook. [MUSIC] [SOUND] [APPLAUSE]
