BUG: Fix integer modulo and division to make integer and float dtypes work similarly for invalid values

[jreback]

closes #3590

This is a numpy oddity that treats them differently.

In [131]: p = DataFrame({ 'first' : [3,4,5,8], 'second' : [0,0,0,3] })

In [132]: p % 0
Out[132]: 
   first  second
0    NaN     NaN
1    NaN     NaN
2    NaN     NaN
3    NaN     NaN

In [133]: p % p
Out[133]: 
   first  second
0      0     NaN
1      0     NaN
2      0     NaN
3      0       0

In [134]: p / p
Out[134]: 
   first    second
0      1       inf
1      1       inf
2      1       inf
3      1  1.000000

In [135]: p / 0
Out[135]: 
   first  second
0    inf     inf
1    inf     inf
2    inf     inf
3    inf     inf

Numpy does this (on integers), floats are as like above

In [3]: x
Out[3]: 
array([[3, 0],
       [4, 0],
       [5, 0],
       [8, 3]])

In [4]: x % 0
Out[4]: 
array([[0, 0],
       [0, 0],
       [0, 0],
       [0, 0]])

In [5]: x % x
Out[5]: 
array([[0, 0],
       [0, 0],
       [0, 0],
       [0, 0]])

In [6]: x / x
Out[6]: 
array([[1, 0],
       [1, 0],
       [1, 0],
       [1, 1]])

In [7]: x / 0
Out[7]: 
array([[0, 0],
       [0, 0],
       [0, 0],
       [0, 0]])

[cpcloud]

@jreback Darn u beat me 2 it :) This is nice, thanks.