Fix compatibility issues with pandas 1.4.3 (#11152)

This PR fixes the two main issues that crop up in pandas 1.4.3 relative to 1.4.2, both around `pd.concat`:
- Columns are now sorted such that integer values come before string values. That is a behavior change that we mimic.
- When multiple objects with identical RangeIndexes are concatenated along axis 1 and sorting is requested, pandas now creates an integer index instead of a RangeIndex. This is not what we want since it increases memory pressure, so those tests have been modified to stop checking the index type and a [pandas issue has been raised](pandas-dev/pandas#47501).

Authors:
  - Vyas Ramasubramani (https://github.com/vyasr)

Approvers:
  - GALI PREM SAGAR (https://github.com/galipremsagar)
  - Nghia Truong (https://github.com/ttnghia)

URL: #11152

Author: vyasr

python/cudf/cudf/core/dataframe.py

@@ -1590,7 +1590,18 @@ def _concat(
                 # include different types that are not comparable.
                 names = sorted(names)
             except TypeError:
-                names = list(names)
+                # For pandas compatibility, we also try to handle the case
+                # where some column names are strings and others are ints. Just
+                # assume that everything that isn't a str is numerical, we
+                # can't sort anything else.
+                try:
+                    str_names = sorted(n for n in names if isinstance(n, str))
+                    non_str_names = sorted(
+                        n for n in names if not isinstance(n, str)
+                    )
+                    names = non_str_names + str_names
+                except TypeError:
+                    names = list(names)
         else:
             names = list(names)
 

python/cudf/cudf/tests/test_concat.py

@@ -803,7 +803,10 @@ def test_concat_join_axis_1(objs, ignore_index, sort, join, axis):
         axis=axis,
     )
 
-    assert_eq(expected, actual, check_index_type=True)
+    # TODO: Remove special handling of check_index_type below
+    # after the following bug from pandas is fixed:
+    # https://github.com/pandas-dev/pandas/issues/47501
+    assert_eq(expected, actual, check_index_type=not (axis == 1 and sort))
 
 
 @pytest.mark.parametrize("ignore_index", [True, False])
@@ -870,7 +873,10 @@ def test_concat_join_one_df(ignore_index, sort, join, axis):
         [gdf1], sort=sort, join=join, ignore_index=ignore_index, axis=axis
     )
 
-    assert_eq(expected, actual, check_index_type=True)
+    # TODO: Remove special handling of check_index_type below
+    # after the following bug from pandas is fixed:
+    # https://github.com/pandas-dev/pandas/issues/47501
+    assert_eq(expected, actual, check_index_type=not (axis == 1 and sort))
 
 
 @pytest.mark.parametrize(
@@ -919,7 +925,10 @@ def test_concat_join_no_overlapping_columns(
         axis=axis,
     )
 
-    assert_eq(expected, actual, check_index_type=True)
+    # TODO: Remove special handling of check_index_type below
+    # after the following bug from pandas is fixed:
+    # https://github.com/pandas-dev/pandas/issues/47501
+    assert_eq(expected, actual, check_index_type=not (axis == 1 and sort))
 
 
 @pytest.mark.parametrize("ignore_index", [False, True])
@@ -1107,13 +1116,14 @@ def test_concat_join_series(ignore_index, sort, join, axis):
         axis=axis,
     )
 
-    # TODO: Remove special handling below
-    # after following bug from pandas is fixed:
+    # TODO: Remove special handling of check_index_type below
+    # after the following bugs from pandas are fixed:
     # https://github.com/pandas-dev/pandas/issues/46675
+    # https://github.com/pandas-dev/pandas/issues/47501
     assert_eq(
         expected,
         actual,
-        check_index_type=False if axis == 1 and join == "outer" else True,
+        check_index_type=(axis == 0),
     )