Used masked table instead of object columns?

[taldcroft]

Currently if one explicitly sets continuity so that auto-fetching of continuity is disabled, then the output columns are all Object type. Instead this could return MaskedColumns with the missing values masked. Maybe this is only low priority since this mode of operation is atypical.

In [33]: states.get_states('2017:300', '2017:302', state_keys=['q1', 'fep_count'], continuity={})
Out[33]:
<Table masked=False length=92>
      datestart              datestop              q1        fep_count trans_keys
      string168             string168            object        object    object
--------------------- --------------------- ---------------- --------- ----------
2017:300:12:00:00.000 2017:300:14:45:15.955             None      None
2017:300:14:45:15.955 2017:300:14:49:15.955     -0.387726852      None         q1
2017:300:14:49:15.955 2017:300:14:54:01.103     -0.387726852      None         q1
2017:300:14:54:01.103 2017:300:14:54:25.103     -0.387726852         0  fep_count
2017:300:14:54:25.103 2017:300:16:42:23.354     -0.387726852         6  fep_count