Merge pull request #8948 from behzadnouri/left-merge

BUG: preserve left frame order in left merge

Author: jreback

doc/source/whatsnew/v0.15.2.txt

@@ -101,6 +101,7 @@ Bug Fixes
 - ``slice`` string method now takes step into account (:issue:`8754`)
 - Bug in ``BlockManager`` where setting values with different type would break block integrity (:issue:`8850`)
 - Bug in ``DatetimeIndex`` when using ``time`` object as key (:issue:`8667`)
+- Bug in ``merge`` where ``how='left'`` and ``sort=False`` would not preserve left frame order (:issue:`7331`)
 - Fix negative step support for label-based slices (:issue:`8753`)
 - Fix: The font size was only set on x axis if vertical or the y axis if horizontal. (:issue:`8765`)
 

pandas/tools/merge.py

@@ -479,8 +479,10 @@ def _get_join_indexers(left_keys, right_keys, sort=False, how='inner'):
         left_group_key, right_group_key, max_groups = \
             _factorize_keys(left_group_key, right_group_key, sort=sort)
 
+    # preserve left frame order if how == 'left' and sort == False
+    kwargs = {'sort':sort} if how == 'left' else {}
     join_func = _join_functions[how]
-    return join_func(left_group_key, right_group_key, max_groups)
+    return join_func(left_group_key, right_group_key, max_groups, **kwargs)
 
 
 class _OrderedMerge(_MergeOperation):

pandas/tools/tests/test_merge.py

@@ -1022,6 +1022,13 @@ def test_left_join_index_multi_match_multiindex(self):
 
         tm.assert_frame_equal(result, expected)
 
+        # GH7331 - maintain left frame order in left merge
+        right.reset_index(inplace=True)
+        right.columns = left.columns[:3].tolist() + right.columns[-1:].tolist()
+        result = merge(left, right, how='left', on=left.columns[:-1].tolist())
+        expected.index = np.arange(len(expected))
+        tm.assert_frame_equal(result, expected)
+
     def test_left_join_index_multi_match(self):
         left = DataFrame([
             ['c', 0],
@@ -1059,6 +1066,11 @@ def test_left_join_index_multi_match(self):
 
         tm.assert_frame_equal(result, expected)
 
+        # GH7331 - maintain left frame order in left merge
+        result = merge(left, right.reset_index(), how='left', on='tag')
+        expected.index = np.arange(len(expected))
+        tm.assert_frame_equal(result, expected)
+
     def test_join_multi_dtypes(self):
 
         # test with multi dtypes in the join index