difference between a= and a(:)=

I find matlab code with assignments

a(:)=... and a(1:N)=...

where in the latter case the right hand side has length N also.

Why they dont write just a=...

Is there a difference?

Also I found the code snipped

if isvector(V)

V = V(:);

what is the sense of this??

Answer 1

[Chunru]

a(:) = 1 will not change the size of a, but assigne 1 to every element of a (assume a has been declared before, such as a = zeros(20,1));

On the other hand, a = 1 (for example) will redefine a (both in size, type and value) so that a becomes 1. Even if a is defined earlier as a = zeros(20, 1), a will be redefined with new rhs.

Answer 2

[KSSV]

a(1:N) will be either row or column vector depedning on a is row or column vector respectively.

a(:) always gives you a column vector.

a = rand(1,10) ;
a(1:10)

ans = 1×10

0.6861 0.9950 0.8572 0.1173 0.9764 0.1016 0.8943 0.4731 0.8291 0.0097

a(:)

ans = 10×1

0.6861 0.9950 0.8572 0.1173 0.9764 0.1016 0.8943 0.4731 0.8291 0.0097

If A is a matrix, A(:) will give a column vector will all columns appended one after another.

You can try yourself instead asking such simple questions.

Answer 3

[Walter Roberson]

Assigning to a(:) or a(1:N) keeps the same "shape" and datatype.

Example:

A1 = zeros(5,1,'uint8');
A2 = zeros(5,1,'uint8');

A1(:) = 1:5;
A2    = 1:5;

After, A1 stays a uint8 column vector, but A2 becomes a double precision row vector.

[Ernst Reißner]

also worth accepting