Back to tutorial index

One dimensional arrays


Topics in this lab

Introduction

Arrays are useful in a variety of situations, including plotting, storing tables of numbers, evaluating mathematical expressions and so on. In this lab, we want to explore ways to create one-dimensional arrays, how to retrieve particular entries of an array, as well as several useful functions for operating on one dimenensional arrays.

clear all
format short

Back to the top

Using square brackets

The simplest way to create an array is to use square brackets to enclose a list of numbers.

v = [1 3 5 7 9]
v =

     1     3     5     7     9

or

v = [1.1 -3.4 1.34e2 sqrt(2) 4^5 1/2]
v =

   1.0e+03 *

    0.0011   -0.0034    0.1340    0.0014    1.0240    0.0005

It is often a good idea to use a comma between individual entries, as in

v = [1.1, -3.4, 1.34e2, sqrt(2), 4^5, 1/2]
v =

   1.0e+03 *

    0.0011   -0.0034    0.1340    0.0014    1.0240    0.0005

so that one can easily distinguish, for example, between

v = [1.1 -3.4]
v =

    1.1000   -3.4000

and

v = [1.1, -3.4]
v =

    1.1000   -3.4000

and not inadvertantly create the scalar value

v = [1.1 - 3.4]
v =

   -2.3000

To get help on the use of the square brackets, use the help keyword punct.

>> help paren
......
[ ]  Brackets are used in forming vectors and matrices.
    [6.9 9.64 SQRT(-1)] is a vector with three elements
    separated by blanks. [6.9, 9.64, sqrt(-1)] is the same
    thing. [1+I 2-I 3] and [1 +I 2 -I 3] are not the same.
    The first has three elements and the second has five.
.......
>> help punct
.......
,   Comma.  The comma is used to separate matrix subscripts
    and arguments to functions.  It is also used to separate
    statements in multi-statement lines. In this situation,
    it may be replaced by a semicolon to suppress printing.

Back to the top

Using the colon operator

In many situations, we want to create arrays of values that follow a particular pattern. One very simple pattern is an array of values equally spaced in some interval. For example, to create an integer array of the values $1,2,3,...10$, we use the colon operator and write

v = 1:10
v =

     1     2     3     4     5     6     7     8     9    10

A more general use of the colon (:) operator allows us to specify the step size we wish to take.

v = 0:2:20
v =

     0     2     4     6     8    10    12    14    16    18    20

is a list of even numbers between 0 and 20. The step size can also be negative. For example,

v = 5:-1:-5
v =

     5     4     3     2     1     0    -1    -2    -3    -4    -5

or

v = 50:-5:0
v =

    50    45    40    35    30    25    20    15    10     5     0

The colon operator automatically fills in entries in an array with equally spaced points. However, we can get some unexpected results if the endpoints we specify are not evenly divided by the stepsize we specify. Consider the following example

v = 0:3:10
v =

     0     3     6     9

We notice that the endpoint 10 does not appear in the array. The reason for this is that 10 is not evenly divided by 3, so is not reached in steps of 3 starting from 0. Rather than shorten the last step size, Matlab only includes those values between the specified endpoints that are even multiples of the step size and shifted by the first endpoint.

Carrying the above example further, we see that it is possible to construct an array that contains only a single element :

v = 10:15:20
v =

    10

In this case, one step already takes us beyond the specified endpoint, and so the vector only contains the starting point 10.

What happens if the step size does not have the same sign as our end value minus our start value? In this case, Matlab will create an empty array. The following example illustrates this.

v = 0:-1:10
v =

   Empty matrix: 1-by-0

Here, 10-0 is positive, but our stepsize -1 is negative, and so an empty array is created. Any time we use a step size of 0, we will also get an empty array. The simplest empty array in Matlab is an empty pair of square brackets ([]).

v = []
v =

     []

To get help on the colon operator, use the help keyword colon.

>> help colon
 :  Colon.
    J:K  is the same as [J, J+1, ..., K].
    J:K  is empty if J > K.
    J:D:K  is the same as [J, J+D, ..., J+m*D] where m = fix((K-J)/D).
    J:D:K  is empty if D == 0, if D > 0 and J > K, or if D < 0 and J < K.
.............

Back to the top

Using the linspace command

Another way to create equally spaced arrays of points is to use the linspace command. Like the colon operator, linspace takes a starting value and and ending value for the desired array. But instead of specifying a stepsize, the linspace command requires the number of entries in an array. For example, to create a list of integers $1,2,3...,10$ that we did above, we would use linspace as follows.

v = linspace(1,10,10)
v =

     1     2     3     4     5     6     7     8     9    10

The first argument is the starting value 1, the second argument is the end value (10), and the third argument is the number of values in the array. To create a list of integers $-3,-2,-1,...,5$, use linspace with these arguments.

v = linspace(-3,5,9)
v =

    -3    -2    -1     0     1     2     3     4     5

If we change the number of entries in the array, without changing the starting and ending values, we will effectively change the stepsize. For example, if we increase the number entries in the above array to 17, we will get

v = linspace(-3,5,17)
v =

  Columns 1 through 7

   -3.0000   -2.5000   -2.0000   -1.5000   -1.0000   -0.5000         0

  Columns 8 through 14

    0.5000    1.0000    1.5000    2.0000    2.5000    3.0000    3.5000

  Columns 15 through 17

    4.0000    4.5000    5.0000

You may be suprised by the fact that we just don't double the number of entries to cut the stepsize in half. But the stepsize that we get using linspace is given by the formula $$ \mbox{stepsize} = \frac{\mbox{ending value} - \mbox{starting value}} {(\mbox{number of values})-1} $$
For the example above, we have $\mbox{stepsize} = (5 - (-3))/16 = 0.5$.

Because the connection between number of points and step size is not as apparent when using the linspace command, it is sometimes easier to specify the number of desired intervals. Then it becomes obvious how to cut the stepsize by increasing the number of intervals.

N = 10;                     % Number of intervals
v = linspace(0,1,N+1)       % stepsize is (1-0)/N = 0.1
v =

  Columns 1 through 7

         0    0.1000    0.2000    0.3000    0.4000    0.5000    0.6000

  Columns 8 through 11

    0.7000    0.8000    0.9000    1.0000

v = linspace(0,1,2*N + 1)   % Stepsize is (1-0)/(2*N) = 0.05
v =

  Columns 1 through 7

         0    0.0500    0.1000    0.1500    0.2000    0.2500    0.3000

  Columns 8 through 14

    0.3500    0.4000    0.4500    0.5000    0.5500    0.6000    0.6500

  Columns 15 through 21

    0.7000    0.7500    0.8000    0.8500    0.9000    0.9500    1.0000

One advantage of linspace over the colon operator is that we can always guarantee that the resulting array will have the starting and ending values that we specify. As we saw above, this isn't always true with the colon operator. This advantage becomes particularly important when we want to use non-integer stepsizes. For example, one could inadvertantly try

v = 0:0.3333:1
v =

         0    0.3333    0.6666    0.9999

when what is really meant is

v = linspace(0,1,4)
v =

         0    0.3333    0.6667    1.0000

Both vectors have four entries, but only the second example evenly divides the interval $[0,1]$ into three equal sized subintervals. Also, the second vector has exactly 1 in its last entry, whereas the first vector has 0.9999 in its last entry.

The linspace command is widely used in plotting because as we will see, it is easy to create an array of function values in a specified interval using linspace.

>> help linspace
linspace Linearly spaced vector.
   linspace(X1, X2) generates a row vector of 100 linearly
   equally spaced points between X1 and X2.
 
   linspace(X1, X2, N) generates N points between X1 and X2.
   For N = 1, linspace returns X2.
.............

Back to the top

Column arrays

One dimensional arrays can be either row or column arrays. All of the arrays we have computed above are row arrays and have a single row, and multiple columns. A column array has a single column and multiple rows. When we used square brackets above, we always created a row array.

row_array = [1,2,3,4,5]
row_array =

     1     2     3     4     5

But we can also use the square brackets to create a column array. To do this, we use the ; (semi-colon) operator.

column_array = [1;2;3;4]
column_array =

     1
     2
     3
     4

Whereas row_array has 1 row and 5 columns, the vector column_array has 5 rows and 1 column.
To create column arrays using the colon operator and linspace, we use the transpose operator, or ' (apostrophe). This swaps the rows and columns of the array. For example,

row_array'
ans =

     1
     2
     3
     4
     5

column_array'
ans =

     1     2     3     4

v = (0:5)'
v =

     0
     1
     2
     3
     4
     5

w = linspace(0,5,6)'
w =

     0
     1
     2
     3
     4
     5

To get help on the the operators above, use the help keyword punct.

>> help punct
;   Semicolon.  The semicolon is used inside brackets to indicate
      the ends of the rows of a matrix.  It is also used after an
      expression or statement to suppress printing.
...........
'   Transpose.   X' is the complex conjugate transpose of X.
      X.' is the non-conjugate transpose.
...........

If we are only working with real arrays, we can ignore the fact that ' is the complex conjugate.

Back to the top

Indexing arrays

Probably the most common use of the colon operator is as a way to retrieve a range of entries in an array. The entries in an array are indexed using integers 1,2,3, to N, the length of the array. To retrieve specific entries, we use the parenthesis operator () with a vector valued argument containing any valid set of indices.
As an example, we create an array of equally spaced values using linspace and retrieve various sets of entries.

v = linspace(-1,1,11)
v =

  Columns 1 through 7

   -1.0000   -0.8000   -0.6000   -0.4000   -0.2000         0    0.2000

  Columns 8 through 11

    0.4000    0.6000    0.8000    1.0000

This array has valid indices 1,2,3, ..., 11. We can retrieve the second entry in v as

v(2)
ans =

   -0.8000

We can retrieve multiple elements of an array all at one time by creating an array of valid index locations. To retrieve every other element in an array of 7 elements, we use the integer sequence

index_set = [1,3,5,7];
v(index_set)
ans =

   -1.0000   -0.6000   -0.2000    0.2000

or simply

v([1,3,5,7])
ans =

   -1.0000   -0.6000   -0.2000    0.2000

To construct longer integer sequences, it is natural to use the colon operator

index_set = 1:2:11;
v(index_set)
ans =

   -1.0000   -0.6000   -0.2000    0.2000    0.6000    1.0000

or simply

v(1:2:11)
ans =

   -1.0000   -0.6000   -0.2000    0.2000    0.6000    1.0000

In both cases, we used an array of integers to specify the desired indices. We can also use the colon operator to retrieve the entries in reverse order,

w = v(10:-1:1)
w =

  Columns 1 through 7

    0.8000    0.6000    0.4000    0.2000         0   -0.2000   -0.4000

  Columns 8 through 10

   -0.6000   -0.8000   -1.0000

In fact we can choose entries in order, even duplicating entries :

w = v([5 7 2 9 9 3 2 2])
w =

  Columns 1 through 7

   -0.2000    0.2000   -0.8000    0.6000    0.6000   -0.6000   -0.8000

  Column 8

   -0.8000

In Matlab, the first entry in an array is always 1. If we try to index the zero-th element of an array, we can expect an error

v(0)
Subscript indices must either be real positive integers or logicals.

If we use an index that is larger than the number of elements that we have, we can also expect an error:

v(12)
Index exceeds matrix dimensions.

Back to the top

Special indexing rules

What if you want to get the last three entries in an array? Since the length of an array is often one of the variables in the script you are writing, you might consider doing the following.

n = length(v);
v([n-2 n-1 n])
ans =

    0.6000    0.8000    1.0000

Here, we used the Matlab length function to find out how long v is.

The above situation arises so often, that Matlab has built into the language a way to get the last entries of a vector by allowing the use of the end keyword inside of arrays. This statement retrieves the last entry in an array

v(end)
ans =

     1

To retrieve the last three entries in an array, you can use the end as if it were the actual variable or constant storing the length of the array. For example,

v([end-2 end-1 end])
ans =

    0.6000    0.8000    1.0000

or simply

v(end-2:end)
ans =

    0.6000    0.8000    1.0000

The colon operator by itself in an array changes the shape of the array from a row vector to a column vector.

v(:)
ans =

   -1.0000
   -0.8000
   -0.6000
   -0.4000
   -0.2000
         0
    0.2000
    0.4000
    0.6000
    0.8000
    1.0000

Example

Suppose we want to compute the midpoint between successive locations along a straight line. We could do the following :

x = [0.1 4.3 5.6 10.9 13.5 16.7 21.9]     % Locations along a line (feet)
x =

    0.1000    4.3000    5.6000   10.9000   13.5000   16.7000   21.9000

midpt = (x(1:end-1) + x(2:end))/2        % Midpoint between each pair of points.
midpt =

    2.2000    4.9500    8.2500   12.2000   15.1000   19.3000

Back to the top

Simple array concatenation

We can always construct a new array by gluing two arrays together. For example, suppose we have two row vectors u and v. We can construct a new array w by glueing, or concatenating u and v together using square brackets.

u = -5:-2
u =

    -5    -4    -3    -2

v = 21:27
v =

    21    22    23    24    25    26    27

w1 = [u v]
w1 =

    -5    -4    -3    -2    21    22    23    24    25    26    27

w2 = [v u]
w2 =

    21    22    23    24    25    26    27    -5    -4    -3    -2

The resulting arrays w1 and w2 each have length equal to the sums of the lengths of u and v.
We can concatenate column arrays as well. In this case, we can stack arrays on top of each other using the semi-colon operator.

x = [1;1;1;1];
y = [2;2;2;2];
z = [x;y]
z =

     1
     1
     1
     1
     2
     2
     2
     2

In a later lab, we will see that it is possible to concatenate column arrays horizontally, or row arrays vertically to obtain two dimensional arrays. Here a preview of what is to come.

A = [x y]
A =

     1     2
     1     2
     1     2
     1     2

Back to the top

Array functions for one-dimensional arrays

Matlab has several functions which operate on one-dimensional arrays. We have already see one such function, the length function, which returns the number of elements in the array.

x = 0:25:1000;
length(x)
ans =

    41

or

x = linspace(0,1000,35);
length(x)
ans =

    35

Many Matlab functions are convenient in that they allow us to avoid the use of the loops that are often required in other languages. A partial list of such functions include min, max, sum, prod, cumsum, cumprod, mean, median, std and sort.

v = 1:10;

We can compute the minimum and maximum values of an array using the min and max functions.

min(v)
ans =

     1

max(v)
ans =

    10

To sum up all the entries in an array, or to muliply all the entries together, use sum and prod.

sum(v)
ans =

    55

prod(v)
ans =

     3628800

The cumulative sum or product of an array x is a second array y whose $n^{th}$ entry is the sum or product of the first n entries of x. The Matlab commands used to produce this array y are cumsum and prodsum.

cumsum(v)
ans =

     1     3     6    10    15    21    28    36    45    55

cumprod(v)
ans =

  Columns 1 through 6

           1           2           6          24         120         720

  Columns 7 through 10

        5040       40320      362880     3628800

In many situations, we want to compute the difference of successive entries of an array. For example, suppose we had an array of locations along a straight line segment and we want to compute the distance between the locations. The diff comes in handy for this.

x = [1.1 4.5 9.1 11.4 21.7]   % locations along a straight line, in feet.
x =

    1.1000    4.5000    9.1000   11.4000   21.7000

diff(x)    % distance between each successive location
ans =

    3.4000    4.6000    2.3000   10.3000

Back to the top

Creating arrays of zeros, ones and random numbers.

We can create arrays of constant values in one of several ways. For example, to create an array of all zeros or all ones, we use the commands zeros or ones.

v = zeros(1,5)
v =

     0     0     0     0     0

w = ones(1,5)
w =

     1     1     1     1     1

To then create an array of all 7s for example, we can either add 7 to an array of all zeros, or multiply an array of all ones by 7.

v = 7 + zeros(1,5)
v =

     7     7     7     7     7

u = 7*ones(1,5)
u =

     7     7     7     7     7

In each the above examples, we passed two arguments to the zeros or ones commands. These commands indicate how many rows and columns each resulting array should have. Here, we created only row or column vectors, but as we will see later, we could just have easily created a two dimensional array

w = ones(3,5)
w =

     1     1     1     1     1
     1     1     1     1     1
     1     1     1     1     1

In many situations, it is useful to have an array of random numbers. To do this, we can use the rand command, which creates an array of random numbers uniformly distributed in the interval [0,1].

format short e
x = rand(1,10)
x =

  Columns 1 through 6

   8.8329e-01   6.7411e-01   4.7052e-01   9.1440e-01   2.5278e-01   5.7975e-01

  Columns 7 through 10

   3.0535e-01   5.2879e-01   5.8160e-01   8.2487e-01

This creates a row vector (i.e. has only 1 row) with 10 entries (i. e. 10 columns) of random numbers.

>> help rand
 rand Uniformly distributed pseudorandom numbers.
    R = rand(N) returns an N-by-N matrix containing pseudorandom values drawn
    from the standard uniform distribution on the open interval(0,1).  rand(M,N)
    or rand([M,N]) returns an M-by-N matrix.  rand(M,N,P,...) or
    rand([M,N,P,...]) returns an M-by-N-by-P-by-... array.  rand returns a
    scalar.  rand(SIZE(A)) returns an array the same size as A.
............

Back to the top

Elementary statistical functions

In this section, we will use array functions that take as input random arrays of numbers.

x = rand(1,100);

The mean, median, standard deviation and variance of an array of numbers can be computed using the mean, median, std and var functions.

mean(x)
ans =

   5.3094e-01

median(x)
ans =

   5.3568e-01

std(x)
ans =

   2.7831e-01

var(x)
ans =

   7.7458e-02

For large enough samples, we expect the mean and median of random integers generated using rand to tend towards 0.5, while the standard deviation tends towards $1/\sqrt{12}$ and variance tends towards $1/12$

format long;
x = rand(1e8,1);
mean(x)
ans =

   0.500013683735540

median(x)
ans =

   0.500023775989154

std(x)
1/sqrt(12)
ans =

   0.288662322475396


ans =

   0.288675134594813

var(x)
1/12
ans =

   0.083325936416889


ans =

   0.083333333333333

Back to the top

Sorting values in an array

We can sort the entries in an array using the sort command.

x = rand(100,1);
x_sorted = sort(x);

To see that we really have a sorted list, we can compare the first entry of x_sorted to the minimum value of x, and the last entry of x_sorted to the maximum value of x.

x_sorted(1)
min(x)
ans =

   0.001733163923346


ans =

   0.001733163923346

x_sorted(end)
max(x)
ans =

   0.993867233183402


ans =

   0.993867233183402

We can use sort to compute the median of a set of numbers. If our array has an odd number of values, the median will be the entry $(n+1)/2$ in the sorted array of length n.

x = rand(101,1);
x_sorted = sort(x);
x_sorted(51)   % Median value : half are below this value; half are above
ans =

   0.535318301218237

This is exactly the value reported by median

median(x)
ans =

   0.535318301218237

For an even number of values, the median is computed as the average of the two middle values in a sorted array.

x = rand(100,1);
x_sorted = sort(x);
(x_sorted(50) + x_sorted(51))/2
ans =

   0.443936398364774

median(x)
ans =

   0.443936398364774

Back to the top

Lab exercises

Use the Matlab functions min, max, sum, prod, cumsum, cumprod, mean, median and std to compute the value of the indicated expressions below. To start, first create an array $x$ of 100 random values using the rand function. For example,
x = rand(100,1);
You will use this array for each of the exercises.

  1. $\displaystyle{\sum_{k=1}^{n} x_k}$
  2. $\displaystyle{\min_{1 \le k \le n} x_k}$
  3. $\displaystyle{\frac{1}{n}\sum_{k=1}^n {x_k}}$
  4. $\displaystyle{n}$
  5. $\displaystyle{y_j = \prod_{k=1}^j {x_k}}$, for $j = 1,2,...,n$
  6. $\displaystyle{\max_{1 \le k \le n} x_k}$
  7. $\displaystyle{\sqrt{\frac{1}{n-1}\sum_{k=1}^n (x_k - \mu)^2}}$, where $\displaystyle{\mu = \frac{1}{n}\sum_{k=1}^n x_k}$
  8. Find the value $\sigma$ in $x$ such that half of the values $x_k$ are less than $\sigma$ and half are greater than $\sigma$.
  9. $\displaystyle{y_j = \sum_{k=1}^j {x_k}}$ for $j = 1,2,...,n$
  10. $\displaystyle{\prod_{k=1}^{n} x_k}$
  1. You are planning a bicycle trip along a 400 mile stretch of a very straight midwestern rural highway and plan to stop each night at a different town. The towns are irregularly spaced, but you have mileage markers for each town, given in the array M :

    $ M = [0, 27, 69, 101, 120, 154, 178, 211, 235, 278, 306, 327, 356, 391, 400] $

    You would like to compute the distances you have to travel each day, as well as other statistics about your trip.

    To start, create an array M containing the mile markers above, and a second array D containing the distances between each of the mile markers and answer the following questions.

    1. What is the shortest distance you will have to bicycle on any day?
    2. What is the longest distance?
    3. What is your average daily distance?
    4. How far will you have to go on day 7?
    5. If you would like to stop each day for lunch at exactly the halfway point for each day's journey, what mileage values should you plug into your GPS to assure that you do not miss lunch?
    6. How can you recover your array M from your array D?
    For these problems, you will want to use diff, mean, min, max and cumsum.

Compare your answers with the solutions.

Back to the top

Get the code

Do you want to try the above code fragments on your own? Download the Matlab script that produces this page here. (lab_4.m)

Powered by MathJax

Published with MATLAB® 8.3