Addtion with 2x1 Vector and 2x3 Matrix

Hi,

In the 2017a documentation, I find that you can "Create an array, A, and add a column vector to it. The vector is treated as though it is a matrix of the same size as A, so that each element in the vector is added to a row in A."

https://www.mathworks.com/help/matlab/ref/plus.html

I am using an older release of Matlab (2015a) and this operation does not work for me. I get the following error:

Error using + Matrix dimensions must agree.

Can anyone tell me how to perform this operation on an older release, and exactly which release was this new functionality available?

Thanks

Answer 1

[Jan]

The new method was introduced in R2016b. See e.g. Loren: arithmetic-expanding. Use bsxfun for older Matlab versions:

A  = rand(2, 3)
b  = rand(2, 1)
R1 = bsxfun(@plus, A, b)  % Matlab <  2016b
R2 = A + b                % Matlab >= 2016b

bsxfun was introduced in 2007a. Before the array way expaned by repmat or ones:

R3 = A + b(:, ones(1, 3))
R4 = A + repmat(b, [1, 3])

The ones-method was processed efficiently by Matlab's JIT acceleration: The timings looked like Matlab does not waste time with creating the expanded array explicitely, but the addition was performed directly. But this is an assumption only: The JIT was never documented exactly.

[EM]

Great thank you for your help. I wish Mathworks would make a note of this and these types of updates in their documentation.

[Jan]

See http://www.mathworks.com/help/matlab/release-notes.html . Almost compact ;-)

[EM]

Thanks for that link! I followed this link and could not find the release note about the change in addition of vectors and matrices of different dimensions though. Where do they list that? Mathematics or Language and Programming?

[James Tursa]

E.g.,

https://www.mathworks.com/help/matlab/matlab_prog/compatible-array-sizes-for-basic-operations.html

[Steven Lord]

In the section for release R2016b it's the first item in the Mathematics subsection. It's called implicit expansion.

[EM]

Thanks for all the help guys! Much appreciated!