Subscripted assignment dimension mismatch.

Hey guys,

I'm pretty new to Matlab so please bear with me :)

I wrote the following code

============================================================================

Result2 = zeros(1, 324012);

for i = 1:324012

    if (Result1(1, i) < 0)

        Result2(1, i)  = dec2bin(typecast(int16(Result1(1,i)),'uint16'));

    else

        Result2(1, i)  = dec2bin(Result1(1,i));

    end

end

============================================================================

And I get the following error

============================================================================

??? Subscripted assignment dimension mismatch.

Error in ==> Untitled4 at 11 Result2(1, i) = dec2bin(Result1(1,i));

============================================================================

What am I doing wrong?

Thanks a lot.

Assaf.

Answer 1

[Matt J]

The output of DEC2BIN is non-scalar, yet you are trying to assign its output to a scalar location Result2(1, i).

What are you trying to accomplish with this code?

Answer 2

[Wayne King]

Because the output of dec2bin() is a char array of a certain number of elements. You do not tell us what Result1 is, but suppose you just had

Result1 = randi([-4 4],3,1);

If I now try to do:

Result1(1) = dec2bin(3);

that will produce the error you see, because I'm trying to assign to one element of Result1 the 1x2 character array: 11

What you can do is to make Result1 a cell array. As an example, I'll just initialize Result1 as a cell array the same dimension as Result2 and just put a single randomly chosen integer from [-4 4] in each element.

     Result2 = zeros(1,324012);
     Result1 = cell(size(Result2));
     [Result1{:}] = deal(randi([-3,3],1,1));
     for ii = 1:324012
        if (Result1{ii} < 0)
            Result2(ii)  = dec2bin(typecast(int16(Result1{ii}),'uint16'));
        else
            Result2(ii)  = dec2bin(Result1{ii});
        end
     end

Answer 3

[assafmalki Malki]

Hey,

Thanks a lot for the help.

Here is the entire code

============================================================================

for Vd=150:2:1000

    n = ceil(2*pi/(A*Vd));

    for k=1:n
        j = j+1;
        Result(j) = cos(A*Vd*k);
    end

end

Result1 = Result * 2^15;

Result2 = zeros(1, 324012);

for i = 1:324012

    if (Result1(1, i) < 0)

        Result2(1, i)  = eval(dec2bin(typecast(int16(Result1(1,i)),'uint16')));

    else

        Result2(1, i)  = eval(dec2bin(Result1(1,i)));

    end

end

============================================================================

As you can see result1 is already a 1x324012 double vector so I don't think I can initialize it as a cell array.

Thanks a lot guys.

Assaf.

[Matt J]

This new version of the code shouldn't produce any errors, now that you're using EVAL. What do you now consider the problem to be?

[Daniyal Awan]

you could still use something like mat2cell() (with proper inputs) to convert your vector into cell-vector of same dimensions. Even though i think using 'eval' is also fine, do you still get error ?

Answer 4

[Matt J]

I'm still not completely sure what you intend to do with this code, but my guess is that your entire code could be replaced by the following 1 line

Result2=dec2bin(uint16(Result1));

[assafmalki Malki]

It didn't help, because it doesn't use 2 complement for the negative numbers, just replacing them with a zero...

Thanks.

Answer 5

[assafmalki Malki]

Hey,

Thanks again.

So this is what I'm trying to do:

I have a vector 1x324012 double named Result, I want a binary representation of this vector. Result is a product of the cosine function so obviously the binary representation of almost all the elements in the vector (except 1) will be zero, that's way I multiply it by 2^15 and try to find the binary representation of that. I also need it to be in 2 complement because cosine can produce negative numbers and also to be outputted to a text file.

If you can tell me a better way to do so, it will be much appreciated.

Thanks a lot.

Assaf.

[Matt J]

Please use Comments to respond to people's Answers and suggestions (like I'm doing now), instead of responding with Answers of your own. This allows readers to distinguish between Answers and non-Answers.

Other than that, have you tried the solutions that people have suggested to you? As far as we know, we've already given you exactly what you want.

Finally, the code you showed us initially is truncating all negative numbers to 0 (by converting them to uint16), so it is not clear why you now say you need 2-s complement.

[assafmalki Malki]

That's way I multiply by 2^15 at the beginning so I wont have numbers such as (0.2323...) for it the to truncate. I need the 2 complement for those numbers (after the multiplication).

Thanks a lot.

Assaf.

[Matt J]

See my latest Answer.

Answer 6

[Matt J]

This converts Result1 to a bit string array containing its two's complement

    Result1=[-3 -2 -1 0 1 2 3];

    neg=Result1<0;
    tc=bitcmp(uint16(-Result1(neg)))+1;

    Result2=uint16(Result1);
    Result2(neg)=tc;

    Result2=dec2bin(Result2),

The output of the above will be

    Result2 =

    1111111111111101
    1111111111111110
    1111111111111111
    0000000000000000
    0000000000000001
    0000000000000010
    0000000000000011

[assafmalki Malki]

Hey,

Looking good, I have one problem with that due, I get as a result one matrix 324012x16 char, I cant even display that, it gives me the following:

Cannot display summaries of variables with more than 524288 elements.

How can I convert it to 324012x1 String?

An another thing for my own knowledge, when I write something like that

tc=bitcmp(uint16(-Result1(neg)))+1;

It only takes into consideration the elements which has '1' in the neg vector?

Thanks a lot.

Assaf.

[Matt J]

Looking good, I have one problem with that due, I get as a result one matrix 324012x16 char, I cant even display that, it gives me the following: Cannot display summaries of variables with more than 524288 elements.

I don't recognize that error message. Are you using the Student Version of MATLAB? In any case, you probably wouldn't want to display such a huge thing anyway. Didn't you plan to save it to a file? You can display some chunk of it by doing things like,

Result2(1:10,:)

How can I convert it to 324012x1 String?

If you want it as a string, you can't make it 324012x1. Each number needs 16 bits and therefore 16 characters.

It only takes into consideration the elements which has '1' in the neg vector?

The "neg" vector was generated as follows neg=(Result1<0); and therefore Result1(neg) will consist only of negative values in Result1. Similarly Result2(neg)=something will only consider elements in those same positions.