FIX: Describe NotImplementedErrors or use AbstractMethodError

This issue fixes pandas-dev#7872 by either describing the reason why a NotImplementedError is raised,
or by throwing an AbstractMethodError in the case of abstract methods (eg: Mixins) that
should be overwritten by the inheriting classes.

This issue is based on discussion of PR pandas-dev#7899 which was never merged.

Author: hoh

pandas/core/base.py

@@ -11,7 +11,7 @@
 import pandas.lib as lib
 from pandas.util.decorators import Appender, cache_readonly
 from pandas.core.strings import StringMethods
-
+from pandas.core.common import AbstractMethodError
 
 _shared_docs = dict()
 _indexops_doc_kwargs = dict(klass='IndexOpsMixin', inplace='',
@@ -32,7 +32,7 @@ class StringMixin(object):
     # Formatting
 
     def __unicode__(self):
-        raise NotImplementedError
+        raise AbstractMethodError(self)
 
     def __str__(self):
         """
@@ -566,4 +566,4 @@ def duplicated(self, take_last=False):
     # abstracts
 
     def _update_inplace(self, result, **kwargs):
-        raise NotImplementedError
+        raise AbstractMethodError(self)

pandas/core/categorical.py

@@ -1166,7 +1166,8 @@ def fillna(self, fill_value=None, method=None, limit=None):
         if fill_value is None:
             fill_value = np.nan
         if limit is not None:
-            raise NotImplementedError
+            raise NotImplementedError("specifying a limit for fillna has not "
+                                      "been implemented yet")
 
         values = self._codes
 

pandas/core/common.py

@@ -39,6 +39,17 @@ class AmbiguousIndexError(PandasError, KeyError):
     pass
 
 
+class AbstractMethodError(NotImplementedError):
+    """Raise this error instead of NotImplementedError for abstract methods
+    while keeping compatibility with Python 2 and Python 3.
+    """
+    def __init__(self, class_instance):
+        self.class_instance = class_instance
+
+    def __str__(self):
+        return "This method must be defined on the concrete class of " \
+               + self.class_instance.__class__.__name__
+
 _POSSIBLY_CAST_DTYPES = set([np.dtype(t).name
                              for t in ['O', 'int8',
                                        'uint8', 'int16', 'uint16', 'int32',

pandas/core/generic.py

@@ -21,7 +21,8 @@
 from pandas.core.common import (isnull, notnull, is_list_like,
                                 _values_from_object, _maybe_promote,
                                 _maybe_box_datetimelike, ABCSeries,
-                                SettingWithCopyError, SettingWithCopyWarning)
+                                SettingWithCopyError, SettingWithCopyWarning,
+                                AbstractMethodError)
 import pandas.core.nanops as nanops
 from pandas.util.decorators import Appender, Substitution, deprecate_kwarg
 from pandas.core import config
@@ -137,7 +138,7 @@ def _init_mgr(self, mgr, axes=None, dtype=None, copy=False):
 
     @property
     def _constructor(self):
-        raise NotImplementedError
+        raise AbstractMethodError(self)
 
     def __unicode__(self):
         # unicode representation based upon iterating over self
@@ -152,7 +153,7 @@ def _local_dir(self):
 
     @property
     def _constructor_sliced(self):
-        raise NotImplementedError
+        raise AbstractMethodError(self)
 
     #----------------------------------------------------------------------
     # Axis
@@ -1100,7 +1101,7 @@ def _iget_item_cache(self, item):
         return lower
 
     def _box_item_values(self, key, values):
-        raise NotImplementedError
+        raise AbstractMethodError(self)
 
     def _maybe_cache_changed(self, item, value):
         """
@@ -3057,7 +3058,8 @@ def first(self, offset):
         """
         from pandas.tseries.frequencies import to_offset
         if not isinstance(self.index, DatetimeIndex):
-            raise NotImplementedError
+            raise NotImplementedError("'first' only supports a DatetimeIndex "
+                                      "index")
 
         if len(self.index) == 0:
             return self
@@ -3091,7 +3093,8 @@ def last(self, offset):
         """
         from pandas.tseries.frequencies import to_offset
         if not isinstance(self.index, DatetimeIndex):
-            raise NotImplementedError
+            raise NotImplementedError("'last' only supports a DatetimeIndex "
+                                      "index")
 
         if len(self.index) == 0:
             return self

pandas/core/groupby.py

@@ -25,7 +25,8 @@
                                notnull, _DATELIKE_DTYPES, is_numeric_dtype,
                                is_timedelta64_dtype, is_datetime64_dtype,
                                is_categorical_dtype, _values_from_object,
-                               is_datetime_or_timedelta_dtype, is_bool_dtype)
+                               is_datetime_or_timedelta_dtype, is_bool_dtype,
+                               AbstractMethodError)
 from pandas.core.config import option_context
 import pandas.lib as lib
 from pandas.lib import Timestamp
@@ -279,7 +280,7 @@ def _set_grouper(self, obj, sort=False):
         return self.grouper
 
     def _get_binner_for_grouping(self, obj):
-        raise NotImplementedError
+        raise AbstractMethodError(self)
 
     @property
     def groups(self):
@@ -670,7 +671,7 @@ def _python_apply_general(self, f):
                                          not_indexed_same=mutated)
 
     def aggregate(self, func, *args, **kwargs):
-        raise NotImplementedError
+        raise AbstractMethodError(self)
 
     @Appender(_agg_doc)
     def agg(self, func, *args, **kwargs):
@@ -680,7 +681,7 @@ def _iterate_slices(self):
         yield self.name, self._selected_obj
 
     def transform(self, func, *args, **kwargs):
-        raise NotImplementedError
+        raise AbstractMethodError(self)
 
     def mean(self):
         """
@@ -1127,7 +1128,7 @@ def _python_agg_general(self, func, *args, **kwargs):
         return self._wrap_aggregated_output(output)
 
     def _wrap_applied_output(self, *args, **kwargs):
-        raise NotImplementedError
+        raise AbstractMethodError(self)
 
     def _concat_objects(self, keys, values, not_indexed_same=False):
         from pandas.tools.merge import concat
@@ -1484,7 +1485,8 @@ def aggregate(self, values, how, axis=0):
                 swapped = True
                 values = values.swapaxes(0, axis)
             if arity > 1:
-                raise NotImplementedError
+                raise NotImplementedError("arity of more than 1 is not "
+                                          "supported for the 'how' argument")
             out_shape = (self.ngroups,) + values.shape[1:]
 
         is_numeric = is_numeric_dtype(values.dtype)
@@ -1556,7 +1558,8 @@ def _aggregate(self, result, counts, values, agg_func, is_numeric):
         comp_ids, _, ngroups = self.group_info
         if values.ndim > 3:
             # punting for now
-            raise NotImplementedError
+            raise NotImplementedError("number of dimensions is currently "
+                                      "limited to 3")
         elif values.ndim > 2:
             for i, chunk in enumerate(values.transpose(2, 0, 1)):
 
@@ -1815,7 +1818,8 @@ def _aggregate(self, result, counts, values, agg_func, is_numeric=True):
 
         if values.ndim > 3:
             # punting for now
-            raise NotImplementedError
+            raise NotImplementedError("number of dimensions is currently "
+                                      "limited to 3")
         elif values.ndim > 2:
             for i, chunk in enumerate(values.transpose(2, 0, 1)):
                 agg_func(result[:, :, i], counts, chunk, self.bins)
@@ -2622,7 +2626,8 @@ def aggregate(self, arg, *args, **kwargs):
             if self._selection is not None:
                 subset = obj
                 if isinstance(subset, DataFrame):
-                    raise NotImplementedError
+                    raise NotImplementedError("Aggregating on a DataFrame is "
+                                              "not supported")
 
                 for fname, agg_how in compat.iteritems(arg):
                     colg = SeriesGroupBy(subset, selection=self._selection,
@@ -2671,7 +2676,7 @@ def _aggregate_multiple_funcs(self, arg):
         from pandas.tools.merge import concat
 
         if self.axis != 0:
-            raise NotImplementedError
+            raise NotImplementedError("axis other than 0 is not supported")
 
         obj = self._obj_with_exclusions
 
@@ -2721,7 +2726,7 @@ def _aggregate_generic(self, func, *args, **kwargs):
         return self._wrap_generic_output(result, obj)
 
     def _wrap_aggregated_output(self, output, names=None):
-        raise NotImplementedError
+        raise AbstractMethodError(self)
 
     def _aggregate_item_by_item(self, func, *args, **kwargs):
         # only for axis==0
@@ -3283,7 +3288,7 @@ def _iterate_slices(self):
                 slice_axis = self._selection_list
             slicer = lambda x: self._selected_obj[x]
         else:
-            raise NotImplementedError
+            raise NotImplementedError("axis other than 0 is not supported")
 
         for val in slice_axis:
             if val in self.exclusions:
@@ -3348,10 +3353,10 @@ def _aggregate_item_by_item(self, func, *args, **kwargs):
             new_axes[self.axis] = self.grouper.result_index
             return Panel._from_axes(result, new_axes)
         else:
-            raise NotImplementedError
+            raise ValueError("axis value must be greater than 0")
 
     def _wrap_aggregated_output(self, output, names=None):
-        raise NotImplementedError
+        raise AbstractMethodError(self)
 
 
 class NDArrayGroupBy(GroupBy):
@@ -3405,7 +3410,7 @@ def _chop(self, sdata, slice_obj):
         return sdata.iloc[slice_obj]
 
     def apply(self, f):
-        raise NotImplementedError
+        raise AbstractMethodError(self)
 
 
 class ArraySplitter(DataSplitter):

pandas/core/internals.py

@@ -294,7 +294,8 @@ def fillna(self, value, limit=None, inplace=False, downcast=None):
         mask = isnull(self.values)
         if limit is not None:
             if self.ndim > 2:
-                raise NotImplementedError
+                raise NotImplementedError("number of dimensions for 'fillna' "
+                                          "is currently limited to 2")
             mask[mask.cumsum(self.ndim-1)>limit]=False
 
         value = self._try_fill(value)
@@ -1681,7 +1682,8 @@ def _slice(self, slicer):
     def fillna(self, value, limit=None, inplace=False, downcast=None):
         # we may need to upcast our fill to match our dtype
         if limit is not None:
-            raise NotImplementedError
+            raise NotImplementedError("specifying a limit for 'fillna' has "
+                                      "not been implemented yet")
 
         values = self.values if inplace else self.values.copy()
         return [self.make_block_same_class(values=values.fillna(fill_value=value,
@@ -1848,7 +1850,8 @@ def fillna(self, value, limit=None,
         value = self._try_fill(value)
         if limit is not None:
             if self.ndim > 2:
-                raise NotImplementedError
+                raise NotImplementedError("number of dimensions for 'fillna' "
+                                          "is currently limited to 2")
             mask[mask.cumsum(self.ndim-1)>limit]=False
 
         np.putmask(values, mask, value)
@@ -2011,7 +2014,8 @@ def interpolate(self, method='pad', axis=0, inplace=False,
     def fillna(self, value, limit=None, inplace=False, downcast=None):
         # we may need to upcast our fill to match our dtype
         if limit is not None:
-            raise NotImplementedError
+            raise NotImplementedError("specifying a limit for 'fillna' has "
+                                      "not been implemented yet")
         if issubclass(self.dtype.type, np.floating):
             value = float(value)
         values = self.values if inplace else self.values.copy()

pandas/core/panelnd.py

@@ -99,7 +99,7 @@ def _combine_with_constructor(self, other, func):
     for f in ['to_frame', 'to_excel', 'to_sparse', 'groupby', 'join', 'filter',
               'dropna', 'shift']:
         def func(self, *args, **kwargs):
-            raise NotImplementedError
+            raise NotImplementedError("this operation is not supported")
         setattr(klass, f, func)
 
     # add the aggregate operations

pandas/core/series.py

@@ -140,7 +140,8 @@ def __init__(self, data=None, index=None, dtype=None, name=None,
                 dtype = self._validate_dtype(dtype)
 
             if isinstance(data, MultiIndex):
-                raise NotImplementedError
+                raise NotImplementedError("initializing a Series from a "
+                                          "MultiIndex is not supported")
             elif isinstance(data, Index):
                 # need to copy to avoid aliasing issues
                 if name is None:

pandas/io/html.py

@@ -19,6 +19,7 @@
                            raise_with_traceback, binary_type)
 from pandas.core import common as com
 from pandas import Series
+from pandas.core.common import AbstractMethodError
 
 _IMPORTS = False
 _HAS_BS4 = False
@@ -229,7 +230,7 @@ def _text_getter(self, obj):
         text : str or unicode
             The text from an individual DOM node.
         """
-        raise NotImplementedError
+        raise AbstractMethodError(self)
 
     def _parse_td(self, obj):
         """Return the td elements from a row element.
@@ -243,7 +244,7 @@ def _parse_td(self, obj):
         columns : list of node-like
             These are the elements of each row, i.e., the columns.
         """
-        raise NotImplementedError
+        raise AbstractMethodError(self)
 
     def _parse_tables(self, doc, match, attrs):
         """Return all tables from the parsed DOM.
@@ -270,7 +271,7 @@ def _parse_tables(self, doc, match, attrs):
         tables : list of node-like
             A list of <table> elements to be parsed into raw data.
         """
-        raise NotImplementedError
+        raise AbstractMethodError(self)
 
     def _parse_tr(self, table):
         """Return the list of row elements from the parsed table element.
@@ -285,7 +286,7 @@ def _parse_tr(self, table):
         rows : list of node-like
             A list row elements of a table, usually <tr> or <th> elements.
         """
-        raise NotImplementedError
+        raise AbstractMethodError(self)
 
     def _parse_thead(self, table):
         """Return the header of a table.
@@ -300,7 +301,7 @@ def _parse_thead(self, table):
         thead : node-like
             A <thead>...</thead> element.
         """
-        raise NotImplementedError
+        raise AbstractMethodError(self)
 
     def _parse_tbody(self, table):
         """Return the body of the table.
@@ -315,7 +316,7 @@ def _parse_tbody(self, table):
         tbody : node-like
             A <tbody>...</tbody> element.
         """
-        raise NotImplementedError
+        raise AbstractMethodError(self)
 
     def _parse_tfoot(self, table):
         """Return the footer of the table if any.
@@ -330,7 +331,7 @@ def _parse_tfoot(self, table):
         tfoot : node-like
             A <tfoot>...</tfoot> element.
         """
-        raise NotImplementedError
+        raise AbstractMethodError(self)
 
     def _build_doc(self):
         """Return a tree-like object that can be used to iterate over the DOM.
@@ -339,7 +340,7 @@ def _build_doc(self):
         -------
         obj : tree-like
         """
-        raise NotImplementedError
+        raise AbstractMethodError(self)
 
     def _build_table(self, table):
         header = self._parse_raw_thead(table)