Is there a way to start indexing with 0 in MATLAB?

a=[4 2 5 2]

can I make this happen?

a(0)=4

Answer 1

[madhan ravi]

Nope not at all index of matlab starts from 1 always!

[IBM watson]

Okay, thanks!

[madhan ravi]

Anytime :)

[Rohit]

intmax("uint8") is 255

now, if I need to look up a uint8 array, I can't since the table is from 1 to 256. Value for 0xFF is in 256 and no way to get to 256 with uint8!?

[Stephen23]

"now, if I need to look up a uint8 array, I can't since the table is from 1 to 256"

You can use any class for the index, it does not need to be UINT8:

A = randi(9,1,999,'uint8')

A = 1×999

5 5 4 4 9 9 6 1 7 6 8 8 8 3 4 1 5 5 7 2 1 8 5 5 8 9 2 2 3 1

A(256) % look up a value in UINT8 array

ans = uint8 7

A(987) % look up other values in UINT8 array

ans = uint8 6

So "to look up a uint8 array" works as expected on this forum. What did you try?

[Bruno Luong]

@Rohit the largest index you can do with MATLAB is

[~,maxsize] = computer

maxsize = 2.8147e+14

This is far larger than 255

[DGM]

If you are working with indexed color images, zero-based indexing is supported, but only when the index array is integer-class.

% an example color table
CT = [0 0 0;
      0.1 0.1 0.1;
      0.2 0.2 0.2;
      0.3 0.3 0.3];

img1 = [1 2 3 4]; % class double
rgb1 = ind2rgb(img1,CT) % indexing starts at 1

rgb1 = 
rgb1(:,:,1) =

         0    0.1000    0.2000    0.3000


rgb1(:,:,2) =

         0    0.1000    0.2000    0.3000


rgb1(:,:,3) =

         0    0.1000    0.2000    0.3000

img2 = uint8([0 1 2 3]); % class uint8
rgb2 = ind2rgb(img2,CT) % indexing starts at 0

rgb2 = 
rgb2(:,:,1) =

         0    0.1000    0.2000    0.3000


rgb2(:,:,2) =

         0    0.1000    0.2000    0.3000


rgb2(:,:,3) =

         0    0.1000    0.2000    0.3000

This behavior should carry over to other tools which support handling of indexed-color images. I don't know of other cases where class-dependent index offsets exist.

[Rohit]

@Stephen23 @Bruno Luong

Below is an assumed example, here data is hard coded to uint8 and not possible to typecast.

@DGM Thank you for letting me know.

>> A = reshape(A.', 1, []); 
>> data = uint8(255); 
>> A(256) 
ans = uint8 66 
>> A(data + 1) 
ans = uint8 109

[Stephen23]

@Rohit: it is unclear why you want to limit yourself to indexing only 256 elements, but I don't see that anything stops you from typecasting. It works on this forum without any problem:

V = rand(1,256); % not UINT8, but has 256 elements.
X = uint8(255); % UINT8 index
V(double(X)+1)

ans = 0.1667

V(256) % just for comparison

ans = 0.1667

[Bruno Luong]

@Rohit

You confuse between uint8 arithmetics (that can never return anything beyond (0:255))

and indexing range.

[Rohit]

@Stephen23 @Bruno Luong Thank you for your inputs. I was under the assumption that typecasting the index would change the the data type of the output which is not the case.

Answer 2

[Bruno Luong]

Yes, but you won't get much support by built-in array stock function

>> M = containers.Map('KeyType','uint32','ValueType','double')

M = 

  Map with properties:

        Count: 0
      KeyType: uint32
    ValueType: double

>> M(0)=3

M = 

  Map with properties:

        Count: 1
      KeyType: uint32
    ValueType: double

>> M(0)

ans =

     3

>> 

[IBM watson]

Great! I will take a look . Thanks

[Walter Roberson]

Also in the File Exchange someone posted a toy class to permit indexing starting from 0.

[IBM watson]

Thanks!

[Walter Roberson]

https://www.mathworks.com/matlabcentral/fileexchange/33301-zero-based-indexable-arrays-oop-exercise

[IBM watson]

Many thanks!!

[Utku Yilmaz]

ZeroBased function (link given above) is very useful. 1 based indexing creates problem when I share my code with software developers using C language. The indexes get messy and it becomes very hard to compare codes.

I always used index values starting from 0, as following. But I think, I will utilize ZeroBased function from now on.

for index_A = 0 : 10
    for index_B = 0 : 10
        value = a_matrix(index_A+1, index_B+1);
    end
end

Answer 3

[Bruno Luong]

One way is to define your onw class with overloading subsref and subsasg

https://www.mathworks.com/help/matlab/matlab_oop/code-patterns-for-subsref-and-subsasgn-methods.html