A Problem with log10

Dear all

I just began to work w Matlab. I have an array with decimal numbers and zeros from which I have to calculate log10 and write a new array, but it doesn't work. When i run this i got error "Attempted to access log10(0.946491); index must be a positive integer or logical." code is as follows

log10=zeros(2030,1354);
for j=1:1354
for i=1:2030
log10(i,j)=R547(i,j)/R531(i,j);
end
end
log10(isnan(log10)) = 0 ; %my array
>> for j=1:1354
for i=1:2030
if log10(i,j)~=0
X(i,j)=log10(double(log10(i,j)));
end
end
end

Anyone please help me

Thank you

[Stephen23]

This is the danger of overwriting inbuilt function names with your own functions or variables. Some common ones that should be avoided:

• log
• i, j
• plot
• sum
• input
• size, length, etc.

Answer 1

[Thorsten]

You overwrite Matlabs log10 function with your log10 matrix; rename it and everything should work fine.

[Alina]

I tried but the answer is the same

U=zeros(2030,1354);
for j=1:1354
for i=1:2030
U(i,j)=R547(i,j)/R531(i,j);
end
end
U(isnan(U)) = 0 ; 
>> for j=1:1354
for i=1:2030
if U(i,j)~=0
X(i,j)=log10(double(U(i,j)));
end
end
end

Attempted to access log10(0.946491); index must be a positive integer or logical.

[Thorsten]

Your log10 matrix is still in the workspace; use

 clear log10

And then have a look which log10 is used now

 which log10

BTW: you can get rid of the for loops:

 U = R547./R341;
 U(isnan(U)) = 0;
 idx = find(U~=0);
 X(idx) = log10(U(idx));

[Alina]

Thank you Thorsten

It works, but some answeres are 0.000000000000000 + 0.000000000000000i where must be 0

[Thorsten]

Use

X = real(X);

[Alina]

Thank you!

Answer 2

[Azzi Abdelmalek]

Try this

clear log10
U=zeros(2030,1354);
for j=1:1354
  for i=1:2030
      U(i,j)=R547(i,j)/R531(i,j);
  end       
end
U(isnan(U)) = 0 ; 
for j=1:1354
  for i=1:2030
      if U(i,j)~=0
             X(i,j)=log10(double(U(i,j)));
      end
  end
end

[Alina]

Thank you!

Answer 3

[Star Strider]

Your code can be reduced to this:

R547 = randi(50,10,15);             % Created Data
R531 = randi(50,10,15);             % Created Data
Log_10 = double(R547./R531);        % Do Element-Wise Division
Log_10(1:5:end,1:5:end) = NaN;      % Created Data (Insert NaN)
X = log10(Log_10);                  % Take Logs
X(isnan(X)) = 0;                    % Set ‘NaN’ Values To Zero

First, please never name your variables the same as built-in MATLAB functions, such as log10. MATLAB will take the most recent definition (as it did in your code) and treat it as a variable rather than a function. I renamed your ‘log10’ array ‘Log_10’ to be sure that you do not ‘overshadow’ the log10 function.

Second, if you need to do element-wise operations, use the dot operators, for instance instead of (/), use (./), (*), use (.*) and (^), use (.^).

Third, the log10 of NaN is NaN, so you only need to test for it and replace it after you do everything else.

[Alina]

Thank you. Now I know it.

[Star Strider]

My pleasure!