deriving new variable from existing column vector using for-loop

Hello Matlab community!

I am a newbie matlab user and trying to be comfortable in using for-loops to derive a new variable.

In my example code below. I designated variable 'c' as my column vector of zeros and ones. I wish to run a for loop for that variable and create a new variable 'data_new' with the following conditions:

** if the value of c is 1 then assign a random variable between 5 to 29 else 0.

**it seems to create those zeros but failed to assign a random variable for the if condition that I stated. here's the error message that I encountered: Unable to perform assignment because the left and right sides have a different number of elements.

a=ones(1,4);
b=zeros(1,4);
c=[a b]'; %column vector 8x1


existing_data=c
for i=1:length(c)
     if (existing_data(i)==0)
         data_new(i)=randi([5,29],8,1)' %%  i wish to generate the same size column vector where the ones == randomly assigned data bet 5-29
     else
         data_new(i)=0
     end
end

Thank you!

Answer 1

[Harry Laing]

Simple error. Your data_new(i)=randi([5,29],8,1)' is the problem. Try putting the line randi([5,29],8,1) ' into the command window and see what happens. You're wanting to assign a single value into data_new but your code tries to assign an entire vector, hence the error.

Try this line instead (change the 8 to a 1):

data_new(i)=randi([5,29],1,1);

As side note, MATLAB will provide you with a warning saying that data_new will change size each iteration. Whilst it may not make much difference with such a small vector, with larger datasets this can cause issues with making the code run much slower. Wherever possible, it is good practice to pre-initiaalise the vector before the loop. For numeric matrices I personally like to create an 'empty' matrix of NaN values before loops like yours. For example, I would write the following into your script before the for loop, but after you create the existing_data variable:

data_new = NaN( size(existing_data) );

[Arturo Jr. Ongkeko]

Thanks, Harry! I will try your suggestion.

Yes, I am actually working with a large dataset but I am trying it first in a small data set, hence, my example. Will surely take note also of your point on pre-initialisation.

[Harry Laing]

randi is only generating 1 value here, between 5 and 29 (the 1,1 part says to create a vector of size 1x1). Each time the for loop passes, a new number will be generated regardless of the previous. The value chosen is already saved, in your data_new variable.

Do you mean you want the random numbers generated to be the same every time you run your code? By definition, that's not random. But to do this I would generate a long list of random numbers in a vector, save the variable, then load it every time you run your code so the 'random' numbers are the same each time.

[Arturo Jr. Ongkeko]

Hi, Harry. Appreciate your reply. I saved my workspace where I created the random values and created another script for data analysis.