indexer silently fails for missing columns

[richjoyce]

df = pd.DataFrame(np.random.randn(6,2), columns=['test_x', 'test_z'])
pv.indexer(df, 'test')

Result: df with ['test_x', 'test_y', 'test_z'] with 'test_y' being NaN
Expected: error complaining no test_y