ENH: raise_assert_detail shows the difference between the columns (#48390)

* show first diff in assert_frame_equal

* reword diff -> first_diff

Author: MarcoGorelli

doc/source/whatsnew/v1.6.0.rst

@@ -29,6 +29,7 @@ enhancement2
 Other enhancements
 ^^^^^^^^^^^^^^^^^^
 - :meth:`Series.add_suffix`, :meth:`DataFrame.add_suffix`, :meth:`Series.add_prefix` and :meth:`DataFrame.add_prefix` support an ``axis`` argument. If ``axis`` is set, the default behaviour of which axis to consider can be overwritten (:issue:`47819`)
+- :func:`assert_frame_equal` now shows the first element where the DataFrames differ, analogously to ``pytest``'s output (:issue:`47910`)
 -
 
 .. ---------------------------------------------------------------------------

pandas/_libs/testing.pyx

@@ -91,6 +91,7 @@ cpdef assert_almost_equal(a, b,
         Py_ssize_t i, na, nb
         double fa, fb
         bint is_unequal = False, a_is_ndarray, b_is_ndarray
+        str first_diff = ''
 
     if lobj is None:
         lobj = a
@@ -159,12 +160,14 @@ cpdef assert_almost_equal(a, b,
             except AssertionError:
                 is_unequal = True
                 diff += 1
+                if not first_diff:
+                    first_diff = f"At positional index {i}, first diff: {a[i]} != {b[i]}"
 
         if is_unequal:
             from pandas._testing import raise_assert_detail
             msg = (f"{obj} values are different "
                    f"({np.round(diff * 100.0 / na, 5)} %)")
-            raise_assert_detail(obj, msg, lobj, robj, index_values=index_values)
+            raise_assert_detail(obj, msg, lobj, robj, first_diff=first_diff, index_values=index_values)
 
         return True
 

pandas/_testing/asserters.py

@@ -639,7 +639,9 @@ def assert_timedelta_array_equal(
         assert_attr_equal("freq", left, right, obj=obj)
 
 
-def raise_assert_detail(obj, message, left, right, diff=None, index_values=None):
+def raise_assert_detail(
+    obj, message, left, right, diff=None, first_diff=None, index_values=None
+):
     __tracebackhide__ = True
 
     msg = f"""{obj} are different
@@ -674,6 +676,9 @@ def raise_assert_detail(obj, message, left, right, diff=None, index_values=None)
     if diff is not None:
         msg += f"\n[diff]: {diff}"
 
+    if first_diff is not None:
+        msg += f"\n{first_diff}"
+
     raise AssertionError(msg)
 
 

pandas/tests/util/test_assert_frame_equal.py

@@ -151,7 +151,8 @@ def test_frame_equal_index_mismatch(check_like, obj_fixture):
 
 {obj_fixture}\\.index values are different \\(33\\.33333 %\\)
 \\[left\\]:  Index\\(\\['a', 'b', 'c'\\], dtype='object'\\)
-\\[right\\]: Index\\(\\['a', 'b', 'd'\\], dtype='object'\\)"""
+\\[right\\]: Index\\(\\['a', 'b', 'd'\\], dtype='object'\\)
+At positional index 2, first diff: c != d"""
 
     df1 = DataFrame({"A": [1, 2, 3], "B": [4, 5, 6]}, index=["a", "b", "c"])
     df2 = DataFrame({"A": [1, 2, 3], "B": [4, 5, 6]}, index=["a", "b", "d"])