pd.save/load does not recreate some dataframe attributes

Ran into this while trying some dropna and merge type operations
on a pickled data frame I loaded back i, It died with an AttributeError
on self._known_consolidated.

df = pd.DataFrame( [[1,2]])
df.save("/tmp/1.pickle")
print df._data._known_consolidated
df2=df.load("/tmp/1.pickle")
print df2._data._known_consolidated

looking at internals.BLockManager.__setstate__ and __getstate__ Here, those
attributes are not saved off.
a quick fix would be to just init to false on load, But is there a legacy reason
why getstate drops everything beyond the 3rd arg or can that be extended?
is there a versioning mechanism?

[wesm]

Ah. That should just be initialized to False in __setstate__. Can you add a test case?

closed via 0b6b472