PERF: Performance improvement for MultiIndexes with a DatetimeIndex level

Author: Mike Kelly

doc/source/v0.15.0.txt

@@ -932,6 +932,7 @@ Performance
 - Performance improvement in writing to sql (``to_sql``) of up to 50% (:issue:`8208`).
 - Performance benchmarking of groupby for large value of ngroups (:issue:`6787`)
 - Performance improvement in ``CustomBusinessDay``, ``CustomBusinessMonth`` (:issue:`8236`)
+- Performance improvement for ``MultiIndex.values`` for multi-level indexes containing datetimes (:issue:`8543`)
 
 
 

pandas/core/index.py

@@ -2937,10 +2937,11 @@ def values(self):
 
         values = []
         for lev, lab in zip(self.levels, self.labels):
-            taken = com.take_1d(lev.values, lab)
+            lev_values = lev.values
             # Need to box timestamps, etc.
             if hasattr(lev, '_box_values'):
-                taken = lev._box_values(taken)
+                lev_values = lev._box_values(lev_values)
+            taken = com.take_1d(lev_values, lab)
             values.append(taken)
 
         self._tuples = lib.fast_zip(values)

vb_suite/index_object.py

@@ -117,3 +117,16 @@
 multiindex_from_product = Benchmark('MultiIndex.from_product(iterables)',
                                     setup, name='multiindex_from_product',
                                     start_date=datetime(2014, 6, 30))
+
+#----------------------------------------------------------------------
+# MultiIndex with DatetimeIndex level
+
+setup = common_setup + """
+level1 = range(1000)
+level2 = date_range(start='1/1/2012', periods=10)
+"""
+
+multiindex_with_datetime_level = \
+    Benchmark("MultiIndex.from_product([level1, level2]).values", setup,
+              name='multiindex_with_datetime_level',
+              start_date=datetime(2014, 10, 11))