BUG: ensure column ordering in HDFStore with dup columns and column where specified

Author: jreback

pandas/io/pytables.py

@@ -2910,9 +2910,7 @@ def create_axes(self, axes, obj, validate=True, nan_rep=None, data_columns=None,
 
         # reindex by our non_index_axes & compute data_columns
         for a in self.non_index_axes:
-            labels = _ensure_index(a[1])
-            if not labels.equals(obj._get_axis(a[0])):
-                obj = obj.reindex_axis(labels, axis=a[0])
+            obj = _reindex_axis(obj, a[0], a[1])
 
         # figure out data_columns and get out blocks
         block_obj = self.get_object(obj).consolidate()
@@ -3000,11 +2998,7 @@ def process_axes(self, obj, columns=None):
 
         # reorder by any non_index_axes & limit to the select columns
         for axis, labels in self.non_index_axes:
-            if columns is not None:
-                labels = Index(labels) & Index(columns)
-            labels = _ensure_index(labels)
-            if not labels.equals(obj._get_axis(axis)):
-                obj = obj.reindex_axis(labels, axis=axis)
+            obj = _reindex_axis(obj, axis, labels, columns)
 
         # apply the selection filters (but keep in the same order)
         if self.selection.filter:
@@ -3683,6 +3677,21 @@ class AppendableNDimTable(AppendablePanelTable):
     obj_type = Panel4D
 
 
+def _reindex_axis(obj, axis, labels, other=None):
+    ax = obj._get_axis(axis)
+    labels = _ensure_index(labels)
+    if other is None and labels.equals(ax):
+        return obj
+
+    labels = _ensure_index(labels.unique())
+    if other is not None:
+        labels = labels & _ensure_index(other)
+    if not labels.equals(ax):
+        slicer = [ slice(None, None) ] * obj.ndim
+        slicer[axis] = labels
+        obj = obj.loc[tuple(slicer)]
+    return obj
+
 def _get_info(info, name):
     """ get/create the info for this name """
     try:

pandas/io/tests/test_pytables.py

@@ -2298,7 +2298,6 @@ def test_wide_table(self):
 
     def test_select_with_dups(self):
 
-
         # single dtypes
         df = DataFrame(np.random.randn(10,4),columns=['A','A','B','B'])
         df.index = date_range('20130101 9:30',periods=10,freq='T')
@@ -2308,6 +2307,10 @@ def test_select_with_dups(self):
             result = store.select('df')
             assert_frame_equal(result,df)
 
+            result = store.select('df',columns=['A'])
+            expected = df.loc[:,['A']]
+            assert_frame_equal(result,expected)
+
         # dups accross dtypes
         df = concat([DataFrame(np.random.randn(10,4),columns=['A','A','B','B']),
                      DataFrame(np.random.randint(0,10,size=20).reshape(10,2),columns=['A','C'])],
@@ -2316,9 +2319,18 @@ def test_select_with_dups(self):
 
         with ensure_clean(self.path) as store:
             store.append('df',df)
+
             result = store.select('df')
             assert_frame_equal(result,df)
 
+            expected = df.loc[:,['A']]
+            result = store.select('df',columns=['A'])
+            assert_frame_equal(result,expected)
+
+            expected = df.loc[:,['B','A']]
+            result = store.select('df',columns=['B','A'])
+            assert_frame_equal(result,expected)
+
     def test_wide_table_dups(self):
         wp = tm.makePanel()
         with ensure_clean(self.path) as store: