Array operations in Matlab?

  1. Array operations in Matlab?????

    Is it possible to apply operations to the rows of a matrix as an array? Such as I have a matix A which is 30 rows by 1 column and I want to treat each row as a vector such that I can solve for the polynomial roots of each row using the roots comand (treating each row as its own equation and the matrix A a system of eqautions)?

    Also would it be possible to take the square roots of each row of a same size matrix. Meaning say I had the same size matrix and I want to take the square root of each row and have the output be a matrix of the same size just with the values of those square roots?
     
  2. jcsd
  3. enigma

    enigma 1,815
    Staff Emeritus
    Science Advisor
    Gold Member

    Hi, Iamfaster :wink:

    Welcome to PF!

    If you need to perform a function on each row of a matrix like

    function(A), where A is a 30x1 vector, I think you'll have to break it up and write a for loop

    Code (Text):

    for i=[1:length(A)]
         function(A(i));
    end
     
    or if A is 30x2 (or more)
    Code (Text):

    for i=[1:length(A)]
         function(A(i,:));
    end
     
    if you want to take the square root, you can find it using the .^ operator, which raises each element to a power as opposed to the ^ which performs matrix multiplication

    Code (Text):

    A.^(1/2)
     
     
  4. Thanks, I am getting closer now but I still cant seem to enter the code in the right form.
    Say I have a matrix

    g1=
    5 9 8
    4 7 6
    1 2 8

    and I want a matrix g2= roots of each row as if each row was a vector consisting of the coeficients of a polynomial. I have tried using code such as

    for i=[1:length(g1)]
    g2(i)=roots(g1(i));
    end

    but I keep getting errors. What am I doing wrong? What code would yield the results I desrire?
     
  5. enigma

    enigma 1,815
    Staff Emeritus
    Science Advisor
    Gold Member

    You're not indexing the entire row.

    g1(1) doesn't have any meaning for a m by n matrix when n>1.

    If you want to reference a single term, you need to put in the entire address lke so:

    g1(1,2)

    to reference the first row, second column.

    If you want to reference the entire row, you need to put in the : operator like so:

    g1(1,:)

    Just using : by itself is "shorthand" for explicitly defining the starting and ending conditions.

    If your matrix g is 4x4

    Code (Text):

    g=[1,2,3,4
         6,7,8,9
        10,11,12,13
        14,15,16,17]
     
    typing

    a=g(2,2:3) will give the result a=[7,8]
    b=g(:,1) will give the result b=[1;6;10;14]
    c=g(3,1:4) will give the result c=[10,11,12,13]
    and
    d=g(3,:) will give the same result as c. d=[10,11,12,13]

    Hope that helps,
     
  6. also, lose the [ and the ]
    like this:

    for i = 1:length(g1)
    g2(i, :) = roots(g1(i, :));
    end
     
  7. enigma

    enigma 1,815
    Staff Emeritus
    Science Advisor
    Gold Member

    Works with or without. I usually leave them in.
     
  8. yes it works because [1:5] evaluates to: 1, 2, 3, 4, 5 (you could say: for i = [1,4,18])
    but you can tell the loop that you want "it" done for the numbers 1 through 5 instead of telling it that you want "it" done for the numbers 1, 2, 3, 4 and 5.

    (I guess you leave them in because of similarity with other programming languages that usually have brackets there, like in C/C++: for(i=1; i<5; i++) )
     
Know someone interested in this topic? Share this thead via email, Google+, Twitter, or Facebook

Have something to add?