I want to shift a part of a row vector to the right then adding a number before the shifted part

assuming I have

a=[1 2 3 4 0 0];
b=[0 0 1 0]

I want 'a' to be

a=[1 2 3 1.5 4 0]

the 1.5 is half the element before

and b indicates when does the shift occur

Answer 1

[Mehmed Saad]

if b contains only 1 shift

ind = find(b);
a(ind+1:end)=circshift(a(ind+1:end),1);
a(ind+1) = a(ind)/2;

a =

1.0000 2.0000 3.0000 1.5000 4.0000 0

[Mohamed Salah]

thanks for your help

but b can contain more than one shift

[Mehmed Saad]

give me an example of more than one shifts

[Mohamed Salah]

a=[1 2 3 4 5 6 0 0 0 0]; %I always have number of zeros on the right equals the number of shifts
b=[1 1 1 1];

So 'a' should look like

a=[1, 0.5, 2, 1, 3, 1.5, 4, 2, 5, 2.5, 6]

'a' can be around 60 element and so is 'b'

[Mehmed Saad]

Instead of padding zeros at the end of vector a use the following strategy

1. Donot pad zeros at the end of a
2. pad zeros in b such that length of a is equal to length of b
3. replace all the zeros in b with NaN
4. repeat a in 2nd dimension for example if a is [1 2] then after repeating it in 2nd dimension it will be [1 2;1 2]
5. multiply 2nd row(repeated) with b and divide it by 2
6. remove all the NaNs and remaining part will be your answer

a = [1 2 3 4 5 6];
b = [1 1 1 1];

b = [b zeros(1,length(a)-length(b))]; %Making a and b equal in length
b(b==0) = nan;% replacing all zeros in b with NaNs
a = repmat(a,2,1);% repeating the vector array across 2nd dimension
a(2,:) = (a(2,:).*b)/2;% multiplying 2nd row with b and dividing it by 2
a=a(~isnan(a))% index elements which are with out NaNs

The value of a will be

a =

    1.0000    0.5000    2.0000    1.0000    3.0000    1.5000    4.0000    2.0000    5.0000    6.0000

similarly for

    a = [ 1 2 3 4 0];
    b = [0 1 0 0];
    

Result will be

 
a =

    1.0000    2.0000    3.0000    1.5000    4.0000         0

Hope this works