ndims behavior in R2016a

I'm running R2016a and don't understand the results I get with ndims. The doc says:

"N = ndims(A) returns the number of dimensions in the array A. The number of dimensions is always greater than or equal to 2."

and also:

 "The number of dimensions in an array is the same as the length of the size vector of the array. In other words, ndims(A) = length(size(A))."

Trying a scalar, vector, matrix and 5-D array, I get:

>> ndims(1) ans = 3

>>length(size(1)) ans = 2

>> ndims(1:4) Index exceeds matrix dimensions.

>> length(size(1:4)) ans = 2

>> ndims(magic(3)) Index exceeds matrix dimensions.

>> length(size(magic(3))) ans =  2

and if I create a five-dimensional array using cat:

>>length(size(cat(5, [1 2; 4 5], [7 8; 3 2]))) ans = 5

but

>> ndims(cat(5, [1 2; 4 5], [7 8; 3 2]))
Index exceeds matrix dimensions.

Am I misusing ndims somehow? It couldn't be broken, could it??

Thanks

Answer 1

[James Tursa]

You have likely inadvertently created a variable called "ndims" that is shadowing the MATLAB function of the same name. Clear that "ndims" variable and try things again. E.g., using R2016a:

>> ndims(1)
ans =
     2
>> length(size(1))
ans =
     2
>> ndims(1:4)
ans =
     2
>> ndims(magic(3))
ans =
     2
>> length(size(magic(3)))
ans =
     2
>> length(size(cat(5, [1 2; 4 5], [7 8; 3 2])))
ans =
     5
>> ndims(cat(5, [1 2; 4 5], [7 8; 3 2]))
ans =
     5

Now, shadowing the MATLAB function "ndims" with a variable of the same name:

>> ndims = 3
ndims =
     3
>> ndims(1)
ans =
     3
>> ndims(1:4)
Index exceeds matrix dimensions. 
>> ndims(magic(3))
Index exceeds matrix dimensions. 
>> clear ndims
>> ndims(1)
ans =
     2
>> ndims(1:4)
ans =
     2
>> ndims(magic(3))
ans =
     2

[Tony Pryse]

HA! How did I do that?? Oh, well ...