This raises an exception ``` df = DataFrame(randn(5, 3), index=labels, columns=['foo', 'bar', 'baz']) df['str'] = 'value' df.ix[::2, 'str'] = nan ```