Testing / infrastructure to support datetime64 in Series, DataFrame

[wesm]

right now:

In [7]: Series(date_range('1/1/2000', '12/30/2000 23:00:00', freq='BM'))
Out[7]: 
0     2000-01-31 00:00:00
1     2000-02-29 00:00:00
2     2000-03-31 00:00:00
3     2000-04-28 00:00:00
4     2000-05-31 00:00:00
5     2000-06-30 00:00:00
6     2000-07-31 00:00:00
7     2000-08-31 00:00:00
8     2000-09-29 00:00:00
9     2000-10-31 00:00:00
10    2000-11-30 00:00:00
11    2000-12-29 00:00:00

In [8]: Series(date_range('1/1/2000', '12/30/2000 23:00:00', freq='BM')).dtype
Out[8]: dtype('datetime64[us]')

In [9]: Series(date_range('1/1/2000', '12/30/2000 23:00:00', freq='BM')).reindex(range(20)
   ...: )
Out[9]: 
0     9.492768e+14
1     9.517824e+14
2     9.544608e+14
3     9.568800e+14
4     9.597312e+14
5     9.623232e+14
6     9.650016e+14
7     9.676800e+14
8     9.701856e+14
9     9.729504e+14
10    9.755424e+14
11    9.780480e+14
12             NaN
13             NaN
14             NaN
15             NaN
16             NaN
17             NaN
18             NaN
19             NaN

similarly:

In [11]: DataFrame({'A' : rng})
Out[11]: 
                  A
0   949276800000000
1   951782400000000
2   954460800000000
3   956880000000000
4   959731200000000
5   962323200000000
6   965001600000000
7   967680000000000
8   970185600000000
9   972950400000000
10  975542400000000
11  978048000000000

In [12]: DataFrame({'A' : rng}).dtypes
Out[12]: A    int64

[wesm]

The basics are there. Will put any further problems (merge/join comes to mind) on separate issues