shifting across blocks results in NaNs if dataframe has multiple blocks (mixed types)

[hhuuggoo]

variant of #10907

In [1]: import pandas as pd

In [2]: df = pd.DataFrame({'a' : [1],'b':[2]})

In [3]: df
Out[3]: 
   a  b
0  1  2

In [4]: df['c'] = 3.5

In [5]: df.shift(axis=1)
Out[5]: 
    a  b   c
0 NaN  1 NaN

[ajcr]

This bug is also highlighted in issue 10539.

[jreback]

right this is a dupe