CLN: move import, add whatsnew entry

doc/source/whatsnew/v0.20.0.txt

@@ -886,3 +886,4 @@ Bug Fixes
 - Bug in ``pd.melt()`` where passing a tuple value for ``value_vars`` caused a ``TypeError`` (:issue:`15348`)
 - Bug in ``.eval()`` which caused multiline evals to fail with local variables not on the first line (:issue:`15342`)
 - Bug in ``pd.read_msgpack`` which did not allow to load dataframe with an index of type ``CategoricalIndex`` (:issue:`15487`)
+- Bug in ``groupby.apply()`` coercing ``object`` series to numeric types, when not all values were numeric (:issue:`15680`)

pandas/core/groupby.py

@@ -11,7 +11,6 @@
     callable, map
 )
 
-import pandas as pd
 from pandas import compat
 from pandas.compat.numpy import function as nv
 from pandas.compat.numpy import _np_version_under1p8
@@ -3426,6 +3425,7 @@ def _decide_output_index(self, output, labels):
 
     def _wrap_applied_output(self, keys, values, not_indexed_same=False):
         from pandas.core.index import _all_indexes_same
+        from pandas.tools.util import to_numeric
 
         if len(keys) == 0:
             return DataFrame(index=keys)
@@ -3569,7 +3569,7 @@ def first_non_None_value(values):
                 so = self._selected_obj
                 if (so.ndim == 2 and so.dtypes.apply(is_datetimelike).any()):
                     result = result.apply(
-                        lambda x: pd.to_numeric(x, errors='ignore'))
+                        lambda x: to_numeric(x, errors='ignore'))
                     date_cols = self._selected_obj.select_dtypes(
                         include=['datetime', 'timedelta']).columns
                     date_cols = date_cols.intersection(result.columns)