How to sum a row of variables during a loop

Hi, dont know if this is possible but I'll ask anyway.

I have a 100*2 matrix. I want to check if each value equals 1 or zero, if it's 1 multiply it by some number and have that saved as some new variable. I then want to sum those variables per row and store those totals in a new 100*1 matrix.

I understand this code doesnt make actual sense but it indicates what im trying to achieve. Can anyone point me in the right direction?

Thanks in advance to anyone for help, it's really appreciated.

val1 = 0;
val2 = 0;
multiplier1 = 10
multiplier2 = 15

for i = 1:100
     if matrix(i,1) == 1
         val1(i)  = matrix(i,1)*multiplier1
     else
     end
     if matrix(i,2) == 1
         val2(i) = matrix(i,2)*multiplier2
     else
     end
     rowtotal(i) = val1(i) +val2(i)
end

[madhan ravi]

It looks trivial, what do you want to do when no of val1 is not equal to val2?

[Stephen]

Hi madhan, thanks for commenting. This example is probably too far from what Im actually trying to achieve, I asked a more complete question in a comment below if you are interested in lending a hand.

Thanks.

Answer 1

[Star Strider]

See if this does what you want:

matrix = randi([0 1], 100, 2);                                  % Create ‘matrix’
multiplier1 = 10;
multiplier2 = 15;

val = bsxfun(@times, matrix, [multiplier1, multiplier2]);
rowtotal = sum(val,2);

Note that you do not have to test for a given element being 0 or 1, since the 0 values will remain 0 and only the 1 values will be multiplied.

[Stephen]

Hi thanks for the asnwer. unfortunately I cant seem to fit this in. I asked a simpler example rather than my own code. Should have included the full code initally my apologies.

Here's a longer (incorrect) example of exactly what im trying if it makes it clearer. Would constructing the initial matrix using a loop and calculating the totals during the loop be a better approach?

Thanks again!!

cyclesnew  = randi(20000, 100, 1); %100*1 array  of cycles between 0 and 20000
replace = randi([0 1], 100,4); % 100* 4 matrix of random 0-1 to indicate replacement
population = [(LPCopen+cyclesnew), (HPCopen+cyclesnew), (HPTopen+cyclesnew),(LPTopen+cyclesnew) ] ; %100*4 matrix of CSN

%indexing to alternate columns in replace + pop
finalpop = replace(:,[1;1]*(1:size(replace,2)));
finalpop(:,1:2:end) = population; 

disp(finalpop)

LPCsale = 0;
HPCsale = 0;
HPTsale = 0;
LPTsale = 0;

cycleprof = 10;


for i = 1:100 
   
    if finalpop(i,2) == 0
    else
        LPCsale(i) = finalpop(i,1)*cycleprof;   
    end
    if finalpop(i,4) == 0
    else
        HPCsale(i) = finalpop(i,3)*cycleprof;   
    end
    if finalpop(i,6) == 0
    else
        HPTsale(i) = finalpop(i,5)*cycleprof;   
    end
    if finalpop(i,8) == 0
    else
        LPTsale(i) = finalpop(i,7)*cycleprof;   
    end
    
    totalsale(i) =(LPCsale(i) + HPCsale(i) + HPTsale(i) + LPTsale(i));
end 

[Star Strider]

I don’t understand.

What are ‘LPCopen’ and the others?

[Stephen]

the 'open' values are user inputs, all set to zero here

[Star Strider]

This is a bit more efficient, and runs without error. You will have to determine if it does what you want:

LPCopen = 0;
HPCopen = 0;
HPTopen = 0;
LPTopen = 0;

cyclesnew  = randi(20000, 100, 1); %100*1 array  of cycles between 0 and 20000
replace = randi([0 1], 100,4); % 100* 4 matrix of random 0-1 to indicate replacement
population = [(LPCopen+cyclesnew), (HPCopen+cyclesnew), (HPTopen+cyclesnew),(LPTopen+cyclesnew) ] ; %100*4 matrix of CSN
%indexing to alternate columns in replace + pop
finalpop = replace(:,[1;1]*(1:size(replace,2)));
finalpop(:,1:2:end) = population; 

cycleprof = 10;

LM = finalpop(:,2:2:end) == 0;

LPCsale = zeros(size(finalpop,1),1);
HPCsale = zeros(size(finalpop,1),1);
LPTsale = zeros(size(finalpop,1),1);
HPTsale = zeros(size(finalpop,1),1);

LPCsale(LM(:,1)) = finalpop(LM(:,1),1)*cycleprof;
HPCsale(LM(:,2)) = finalpop(LM(:,2),1)*cycleprof;
LPTsale(LM(:,3)) = finalpop(LM(:,3),1)*cycleprof;
HPTsale(LM(:,4)) = finalpop(LM(:,4),1)*cycleprof;

totalsale = LPCsale + HPCsale + HPTsale + LPTsale;

This uses ‘logical indexing’ to eliminate the loops and the if block tests. Otherwise, it is essentially your original code.

[Stephen]

Wow thank you so much this is perfect!

I must read up on logical indexing.

Thanks again I really appreciate you taking the time to put this together, you're a life saver!!

[Star Strider]

As always, my pleasure!