fixes 12558

Author: Nick Eubank

doc/source/whatsnew/v0.18.1.txt

@@ -43,3 +43,4 @@ Performance Improvements
 
 Bug Fixes
 ~~~~~~~~~
+- Bug in ``value_counts`` where normalizes over all observations including missing even when ``dropna=True`` (:issue:`12558`)

pandas/core/algorithms.py

@@ -269,6 +269,7 @@ def value_counts(values, sort=True, ascending=False, normalize=False,
     from pandas.core.series import Series
     from pandas.tools.tile import cut
     from pandas import Index, PeriodIndex, DatetimeIndex
+    from pandas.core.common import notnull
 
     name = getattr(values, 'name', None)
     values = Series(values).values
@@ -342,7 +343,10 @@ def value_counts(values, sort=True, ascending=False, normalize=False,
         result = result.sort_values(ascending=ascending)
 
     if normalize:
-        result = result / float(values.size)
+        if dropna:
+            result = result / float(values[notnull(values)].size)
+        else:
+            result = result / float(values.size)
 
     return result
 

pandas/tests/test_algos.py

@@ -517,6 +517,12 @@ def test_dropna(self):
                 pd.Series([10.3, 5., 5., None]).value_counts(dropna=False),
                 pd.Series([2, 1, 1], index=[5., 10.3, np.nan]))
 
+    def test_dropna_normalize(self):
+        # Issue 12558
+        tm.assert_series_equal(
+            pd.Series([ 5.,10.3,10.3,10.3,np.nan]).value_counts(dropna=True, normalize=True),
+            pd.Series([0.75, 0.25], index=[10.3, 5.]))
+
 
 class GroupVarTestMixin(object):