factorial function won't take a cell argument independently of it's class (double, integer, etc)

i can't calculate the factorial of a cell element (being it an array), though it is double (checked it with class funct.) and factorial doing fine with a double array... and neither it works using int16 or uint16 conversions on the cell element (called bins {1}) it simply states that "N must be a matrix of non-negative integers." and i do give it such thing.

Answer 1

[Stephen23]

This is because of your MATLAB version. Here using MATLAB 2010b:

>> factorial(uint16([3.4 2.8]))
??? Error using ==> factorial at 17
N must be a matrix of non-negative integers.

>> factorial(round([3.4 2.8]))
ans =
     6     6

The second line of code actually throws an error for this case:

 ~isa(N,'double')

which you would have found by simply clicking on the error message link.

[ethan1987]

ok thx!

Answer 2

[Walter Roberson]

Could you confirm that you are calling factorial(bins{1}) ? If so what is class(bins{1})

[ethan1987]

yes that is what i said...the factorial of a cell element called bins{1}

Answer 3

[Image Analyst]

This works just fine:

% Make a 1-by-10 vector of random integers in the range 1-9
dblBins = randi(9, 1,10)
% Stick into a 1 by 2 cell array.
bins = {dblBins, dblBins}
% Extract the contents of the first cell and take it's factorial
f = factorial(bins{1})

Chances are you have non-integers in there or negative numbers.

[John D'Errico]

Very often, somebody sees something like 1.000, and assumes that it is indeed the integer 1.

format short g
1.000000001
ans =
          1

Just because it looks like an integer does not mean it is one.

[Image Analyst]

Good point John - we see that a lot. If they are intended to be integers they can cast to integers with round() or int32().

[ethan1987]

But i said i checked the class with class function

[John D'Errico]

And what did class tell you? For example:

X = {{ 1 }};
class(X{1})
ans =
cell

factorial(X{1})
Error using factorial (line 20)
N must be an array of real non-negative integers.

X{1} above is a cell itself, despite the fact that it contains an integer. X{1} is NOT an integer.

[ethan1987]

there are no negatives because the numbers are bins of an histogram and their positions are times (positive reals)....also i had the problem of them not being integer, used the round function to no solution and then tried int16 uint16 and the posted error occurs...bins{1} is double as i posted and factorial of a double do work...but i don't know what's happening, factorial(uint16(bins{1})) MUST work...uint are unsigned integers!

[ethan1987]

i do class(bins{1}) and matlab yields DOUBLE

[ethan1987]

sorry, i must say that factorial(round(bins{1})) do work but i gives me a number*array where array have the first few elements 0.000000.... and the inverse of those is Inf ...and thats my problem. factorial(uint16(bins{1})) directly does not work at all

[ethan1987]

also format long gives 0.0000....and so the non zero digit is to far away i guess? is there someway to actually see it? to make sure it is really non zero

[John D'Errico]

Why not attach the array as it is? (Attach it to your next comment as a .mat file.) Then we can see the variable, read it into MATLAB ourselves, and then help you to resolve your problem.

[ethan1987]

ok here it is

[Walter Roberson]

Give the command

format long g

and then print out your values again.

[Image Analyst]

s = load('bins.mat')
b = s.bins{1}

b =
  Columns 1 through 5
                      9.35                     10.05                     10.75                     11.45                     12.15
  Columns 6 through 10
                     12.85                     13.55                     14.25                     14.95                     15.65
  Columns 11 through 15
                     16.35                     17.05                     17.75                     18.45                     19.15
  Columns 16 through 20
                     19.85                     20.55                     21.25                     21.95                     22.65
  Columns 21 through 25
                     23.35                     24.05                     24.75                     25.45                     26.15
  Columns 26 through 30
                     26.85                     27.55                     28.25                     28.95                     29.65
  Columns 31 through 35
                     30.35                     31.05                     31.75                     32.45                     33.15
  Columns 36 through 40
                     33.85                     34.55                     35.25                     35.95                     36.65
  Columns 41 through 45
                     37.35                     38.05                     38.75                     39.45                     40.15
  Columns 46 through 50
                     40.85                     41.55                     42.25                     42.95                     43.65

Is there something in there that does not look like integers??? Why do you think 9.35 factorial is a valid thing to do?

[Walter Roberson]

If you have the symbolic toolbox, then replace

factorial(bins{1})

with

gamma(bins{1}+1)

gamma is the generalization of factorial.

Note: gamma has a singularity at every negative integer.

[ethan1987]

Image analyst: I didn't use factorial on bins as it is . i repeatedly said that i tried to round them or convert into integers!...also factorial of fractionals do exists even though i'm not doing such thing

[ethan1987]

Nevermind, i realized that round(bins{1}) is not null and therefore is not the cause on the Inf in my script calculations, thx anyway...But i still can't understand why factorial(uint16(bins{1})) won't work...factorial(uint16([3.4 2.8])) wont work either...what's the problem there??

[Image Analyst]

This works fine for me:

factorial(round(b))

[John D'Errico]

No problem in the least here.

factorial(uint16([3.4 2.8]))
ans =
      6      6

Note that the result is a uint16 number. So it is trivial to generate an overflow.

factorial(uint16(9.3))
ans =
  65535

And...

factorial(uint16(bins{1})) 
ans =
  Columns 1 through 21
65535  65535  65535  65535  65535  65535  65535  65535  65535  65535  65535  65535  65535  65535  65535  65535  65535  65535  65535  65535  65535
Columns 22 through 42
65535  65535  65535  65535  65535  65535  65535  65535  65535  65535  65535  65535  65535  65535  65535  65535  65535  65535  65535  65535  65535
Columns 43 through 50
65535  65535  65535  65535  65535  65535  65535  65535

Perhaps you need to tell us which release of MATLAB you are trying this with?

[ethan1987]

oh, of course...i didn't get to that part and realized that, hehe... I'm using matlab 7.12 R2011a