What does 2nd argument mean in numel function?

Earlier version of matlab has some extra arguments for the function numel , which said, it gives the number of subscripted elements. What does it mean? I am not able to figure it out. Even the current version supports those extra arguments , but doesn't show it in Documentation. Please Explain what it is.

a = rand(5);
numel(a , 1 , 2 , 3) % it always gives 1 as output, no matter what I try
numel(a , [1 ,2])    % Gives the length of 2nd argument

[Adam]

subsref and subassign both use numel so maybe they make use of the further arguments in a way that is not generally useful for normal code.

I don't generally over-ride either of those two functions though in my classes as they can get complicated so I have never really looked into it.

Answer 1

[Guillaume]

if you do

numel(a, 1:10, 1:10)

it returns 100. It looks like it returns the product of the number of elements of all the arguments starting from the second one and actually ignores the first.

I don't see how that is useful.

[Gautam]

yeah , getting it

numel(a , [2 , 3 , 6 , 8] , [3 , 2 , 9] , [1 ,2])

ans =

    24

But , still what sense does it makes? There must be some usage of it. Thanks for the reply

[Guillaume]

According to the earliest documentation I can find matlab used to call numel for expressions such as A{i1, i2, i3} that generate comma-separated lists, probably to know how many list elements to allocate.

I don't think it does anymore, and it looks like it's broken now as it completely ignores the size of the first argument. I would expect in the past, it would have thrown an error if the indices were out of bounds.