Support for Unsigned 64 INT in datatable

Currently there is no support for unsigned64 Int in datatable
If the unsigned int 64 is imported from pandas it is registered as float64 in datatable
```
>>> x = pd.DataFrame({"A":[9999716810958341233,1231,234523452345],"dd":[1,2,3]}
>>> x.dtypes
A uint64
dd int64
>>> y = dt.Frame(x)
>>>> y
float64 int
--- -------------- ----
0 9.99972e+18 1
1 1231 2
2 2.34523e+11 3
```
To see how fread reads in this data:
```
>>> dt.fread("A\n99997168109583412344\n")
str32
--- --------------------
0 99997168109583412344
```