API: select_dtypes · Issue #7316 · pandas-dev/pandas (original) (raw)
maybe something like this
df.get_for_dtypes(types)
types could be a list of string/dtypes or a dtype/string
e.g.
'float64'
['float']
['float32','float64','object']
['float',np.dtype('bool')]
['numeric']
['datetime','timedelta']
are possible values
pretty easy to implement (just translate the types and filter on the blocks, ala _get_numeric_data)