Iterating through a dataset and creating a vector with according values

So I have a 10 000 X 1 vector (let's call it vector A) of different values ranging from -1 to 1. I want to create another 10 000 X 1 vector (let's call it vector B) that has values corresponding to the ones in the first vector. I want to create a loop that does this:

Everytime a value in vector A is bigger than 0.5, add the value 3 to vector B with the same corresponding position.

Everytime a value in vector A is smaller than -0.5, add the value 1 to vector B with the same corresponding position.

Everytime a value in vector A is between -0.5 and 0.5, add the value 2 to vector B with the same corresponding.

In that case, the resulting matrix should look like this:

[A]     [B]
-0.9    1
-0.3    2
0.6     3
-0.2    2
-0.7    1

I know I need to use a loop but I struggle with this. Thank you so much for your help.

[Dyuman Joshi]

What about border values i.e. 0.5 and -0.5, where value will you assign to corresponding B elements?

[Camille Godin]

0.5 should be included in 3 and -0.5 should be included with 1

Answer 1

[dpb]

Rarely need loops for such things as this with MATLAB; either logical addressing or table lookup almost always comes to the rescue -- the latter (using the builtin interpolation routine) solution here would be

A=[-0.9;-0.3;0.6;-0.2;-0.7];
B=interp1([-1,-0.5,0.5,1],[1,2,3,3],A,'previous')

B = 5×1

1 2 3 2 1

Alternatively, another lookup solution, with less overhead...

B=discretize(A,[-1,-0.5,0.5,1])

B = 5×1

1 2 3 2 1

[Camille Godin]

Thank you, it does work with the interpolation routine.

Now, following the same logic, how can I create another vector that has string values in them instead of 1 2 and 3? If the number is 1, I want the string to be 'GT', if the number is 2, I want the string to be 'IN' and if the number is 3, I want the string to be 'ST'.

[Dyuman Joshi]

@dpb, your answers will not give the answer as expected by OP

A=[-0.9;-0.3;0.6;-0.2;-0.5;0.5];
%border values
% -0.5 =>1 and 0.5 => 3
B=interp1([-1,-0.5,0.5,1],[1,2,3,3],A,'previous')

B = 6×1

1 2 3 2 2 3

B=discretize(A,[-1,-0.5,0.5,1])

B = 6×1

1 2 3 2 2 3

Direct logical indexing might be a good option

C=1*(A<=-0.5)+2*(A>-0.5&A<0.5)+3*(A>=0.5)

C = 6×1

1 2 3 2 1 3

[Dyuman Joshi]

@Camille Godin,

After you get the values, you can get the values as such

str={'GT';'IN';'ST'};
A=[-0.9;-0.3;0.6;-0.2;-0.5;0.5];
B=interp1([-1,-0.5,0.5,1],[1,2,3,3],A,'previous');
%cell array
C=str(B)

C = 6×1 cell array
    {'GT'}
    {'IN'}
    {'ST'}
    {'IN'}
    {'IN'}
    {'ST'}

%char array
D=cell2mat(C)

D = 6×2 char array
    'GT'
    'IN'
    'ST'
    'IN'
    'IN'
    'ST'

[dpb]

If the border edge cases are that important, use bkpt-eps(bkpt) to add the necessary granularity in either solution.

There are multiple other answers I've posted (at least one pretty recently) that illustrate...