PERF: Utilize mixed dtype information in df.count()

Author: qwhelan

doc/source/whatsnew/v0.16.0.txt

@@ -48,6 +48,7 @@ Performance
 - Fixed a performance regression for ``.loc`` indexing with an array or list-like (:issue:`9126`:).
 - Performance improvements in ``MultiIndex.duplicated`` by working with labels instead of values (:issue:`9125`)
 - Improved the speed of `nunique` by calling `unique` instead of `value_counts` (:issue:`9129`, :issue:`7771`)
+- Performance improvement of up to 10x in ``DataFrame.count`` and ``DataFrame.dropna`` by taking advantage of homogeneous/heterogeneous dtypes appropriately (:issue:`9136`)
 
 
 Bug Fixes

pandas/core/frame.py

@@ -4094,11 +4094,11 @@ def count(self, axis=0, level=None, numeric_only=False):
         if len(frame._get_axis(axis)) == 0:
             result = Series(0, index=frame._get_agg_axis(axis))
         else:
-            if axis == 1:
-                counts = notnull(frame.values).sum(1)
-                result = Series(counts, index=frame._get_agg_axis(axis))
-            else:
+            if frame._is_mixed_type:
                 result = notnull(frame).sum(axis=axis)
+            else:
+                counts = notnull(frame.values).sum(axis=axis)
+                result = Series(counts, index=frame._get_agg_axis(axis))
 
         return result.astype('int64')
 

vb_suite/frame_methods.py

@@ -290,30 +290,43 @@ def f(K=100):
                            start_date=datetime(2012,1,1))
 
 ## dropna
-setup = common_setup + """
+dropna_setup = common_setup + """
 data = np.random.randn(10000, 1000)
 df = DataFrame(data)
 df.ix[50:1000,20:50] = np.nan
 df.ix[2000:3000] = np.nan
 df.ix[:,60:70] = np.nan
 """
-frame_dropna_axis0_any  = Benchmark('df.dropna(how="any",axis=0)', setup,
+frame_dropna_axis0_any  = Benchmark('df.dropna(how="any",axis=0)', dropna_setup,
                                      start_date=datetime(2012,1,1))
-frame_dropna_axis0_all  = Benchmark('df.dropna(how="all",axis=0)', setup,
+frame_dropna_axis0_all  = Benchmark('df.dropna(how="all",axis=0)', dropna_setup,
                                      start_date=datetime(2012,1,1))
 
-setup = common_setup + """
+frame_dropna_axis1_any  = Benchmark('df.dropna(how="any",axis=1)', dropna_setup,
+                                    start_date=datetime(2012,1,1))
+
+frame_dropna_axis1_all  = Benchmark('df.dropna(how="all",axis=1)', dropna_setup,
+                                    start_date=datetime(2012,1,1))
+
+# dropna on mixed dtypes
+dropna_mixed_setup = common_setup + """
 data = np.random.randn(10000, 1000)
 df = DataFrame(data)
 df.ix[50:1000,20:50] = np.nan
 df.ix[2000:3000] = np.nan
 df.ix[:,60:70] = np.nan
+df['foo'] = 'bar'
 """
-frame_dropna_axis1_any  = Benchmark('df.dropna(how="any",axis=1)', setup,
-                                    start_date=datetime(2012,1,1))
+frame_dropna_axis0_any_mixed_dtypes  = Benchmark('df.dropna(how="any",axis=0)', dropna_mixed_setup,
+                                                 start_date=datetime(2012,1,1))
+frame_dropna_axis0_all_mixed_dtypes  = Benchmark('df.dropna(how="all",axis=0)', dropna_mixed_setup,
+                                                 start_date=datetime(2012,1,1))
 
-frame_dropna_axis1_all  = Benchmark('df.dropna(how="all",axis=1)', setup,
-                                    start_date=datetime(2012,1,1))
+frame_dropna_axis1_any_mixed_dtypes  = Benchmark('df.dropna(how="any",axis=1)', dropna_mixed_setup,
+                                                 start_date=datetime(2012,1,1))
+
+frame_dropna_axis1_all_mixed_dtypes  = Benchmark('df.dropna(how="all",axis=1)', dropna_mixed_setup,
+                                                 start_date=datetime(2012,1,1))
 
 
 #----------------------------------------------------------------------