BUG: correctly handle py2.6 datetime64 in groupers

Author: jreback

pandas/core/common.py

@@ -2009,6 +2009,14 @@ def needs_i8_conversion(arr_or_dtype):
             is_timedelta64_dtype(arr_or_dtype))
 
 
+def is_numeric_dtype(arr_or_dtype):
+    if isinstance(arr_or_dtype, np.dtype):
+        tipo = arr_or_dtype.type
+    else:
+        tipo = arr_or_dtype.dtype.type
+    return (issubclass(tipo, (np.number, np.bool_))
+            and not issubclass(tipo, (np.datetime64, np.timedelta64)))
+
 def is_float_dtype(arr_or_dtype):
     if isinstance(arr_or_dtype, np.dtype):
         tipo = arr_or_dtype.type

pandas/core/groupby.py

@@ -19,7 +19,8 @@
 import pandas.core.algorithms as algos
 import pandas.core.common as com
 from pandas.core.common import(_possibly_downcast_to_dtype, isnull,
-                               notnull, _DATELIKE_DTYPES)
+                               notnull, _DATELIKE_DTYPES, is_numeric_dtype,
+                               is_timedelta64_dtype, is_datetime64_dtype)
 
 import pandas.lib as lib
 from pandas.lib import Timestamp
@@ -260,9 +261,13 @@ def indices(self):
 
     def _get_index(self, name):
         """ safe get index """
-        if isinstance(name, Timestamp):
-            name = name.value
-        return self.indices[name]
+        try:
+            return self.indices[name]
+        except:
+            if isinstance(name, Timestamp):
+                name = name.value
+                return self.indices[name]
+            raise
 
     @property
     def name(self):
@@ -683,7 +688,7 @@ def _try_cast(self, result, obj):
     def _cython_agg_general(self, how, numeric_only=True):
         output = {}
         for name, obj in self._iterate_slices():
-            is_numeric = _is_numeric_dtype(obj.dtype)
+            is_numeric = is_numeric_dtype(obj.dtype)
             if numeric_only and not is_numeric:
                 continue
 
@@ -721,7 +726,7 @@ def _python_agg_general(self, func, *args, **kwargs):
 
                 # since we are masking, make sure that we have a float object
                 values = result
-                if _is_numeric_dtype(values.dtype):
+                if is_numeric_dtype(values.dtype):
                     values = com.ensure_float(values)
 
                 output[name] = self._try_cast(values[mask], result)
@@ -1087,7 +1092,7 @@ def aggregate(self, values, how, axis=0):
                 raise NotImplementedError
             out_shape = (self.ngroups,) + values.shape[1:]
 
-        if _is_numeric_dtype(values.dtype):
+        if is_numeric_dtype(values.dtype):
             values = com.ensure_float(values)
             is_numeric = True
         else:
@@ -1481,6 +1486,15 @@ def __init__(self, index, grouper=None, name=None, level=None,
                     self.grouper = None  # Try for sanity
                     raise AssertionError(errmsg)
 
+        # if we have a date/time-like grouper, make sure that we have Timestamps like
+        if getattr(self.grouper,'dtype',None) is not None:
+            if is_datetime64_dtype(self.grouper):
+                from pandas import to_datetime
+                self.grouper = to_datetime(self.grouper)
+            elif is_timedelta64_dtype(self.grouper):
+                from pandas import to_timedelta
+                self.grouper = to_timedelta(self.grouper)
+
     def __repr__(self):
         return 'Grouping(%s)' % self.name
 
@@ -1928,7 +1942,7 @@ def _cython_agg_blocks(self, how, numeric_only=True):
         for block in data.blocks:
             values = block.values
 
-            is_numeric = _is_numeric_dtype(values.dtype)
+            is_numeric = is_numeric_dtype(values.dtype)
 
             if numeric_only and not is_numeric:
                 continue
@@ -2980,12 +2994,6 @@ def _reorder_by_uniques(uniques, labels):
 }
 
 
-def _is_numeric_dtype(dt):
-    typ = dt.type
-    return (issubclass(typ, (np.number, np.bool_))
-            and not issubclass(typ, (np.datetime64, np.timedelta64)))
-
-
 def _intercept_function(func):
     return _func_table.get(func, func)