Address PR feedback

Author: edvilme

Lib/statistics.py

@@ -1206,10 +1206,13 @@ def quantiles(data, *, n=4, method='exclusive'):
         result = []
         for i in range(1, n):
             j, delta = divmod(i * m, n)
-            if (data[j] == data[j + 1]) or not delta:
-                interpolated = data[j] / 1
+            a, b = data[j], data[j + 1]
+            if a == b and type(a) is type(b):
+                interpolated = a / 1
+            elif 2 * delta <= n:
+                interpolated = a + (b - a) * delta / n
             else:
-                interpolated = (data[j] * (n - delta) + data[j + 1] * delta) / n
+                interpolated = b - (b - a) * (n - delta) / n
             result.append(interpolated)
         return result
 
@@ -1220,13 +1223,13 @@ def quantiles(data, *, n=4, method='exclusive'):
             j = i * m // n                               # rescale i to m/n
             j = 1 if j < 1 else ld-1 if j > ld-1 else j  # clamp to 1 .. ld-1
             delta = i*m - j*n                            # exact integer math
-            # When the endpoints are equal or delta is zero, avoid
-            # the interpolation formula which can be off by 1 ULP
-            # due to floating-point rounding
-            if (data[j - 1] == data[j]) or not delta:
-                interpolated = data[j - 1] / 1
+            a, b = data[j - 1], data[j]
+            if a == b and type(a) is type(b):
+                interpolated = a / 1
+            elif 2 * delta <= n:
+                interpolated = a + (b - a) * delta / n
             else:
-                interpolated = (data[j - 1] * (n - delta) + data[j] * delta) / n
+                interpolated = b - (b - a) * (n - delta) / n
             result.append(interpolated)
         return result
 

Lib/test/test_statistics.py

@@ -2657,14 +2657,31 @@ def test_monotonic_with_duplicate_floats(self):
                   1e-300,             # small magnitude
                   float.fromhex('0x1.fffffffffffffp+1023'),  # near max float
                   sys.float_info.min, # smallest normal
+                  float('inf'),
+                  float('-inf'),
                   ):
             for method in ('exclusive', 'inclusive'):
                 for n in range(2, 20):
-                    result = quantiles([x, x], n=n, method=method)
-                    self.assertEqual(result, sorted(result),
-                                     msg=f'x={x}, n={n}, method={method!r}')
-                    self.assertTrue(all(v == x for v in result),
-                                    msg=f'x={x}, n={n}, method={method!r}')
+                    with self.subTest(x=x, n=n, method=method):
+                        result = quantiles([x, x], n=n, method=method)
+                        self.assertEqual(result, sorted(result))
+                        self.assertTrue(all(v == x for v in result))
+
+        result = quantiles([0.09999999999999999, 0.1, 0.1],
+                           n=9, method='inclusive')
+        self.assertEqual(result, sorted(result))
+
+    def test_mixed_types(self):
+        data = [Fraction(1, 2), 0.5, 2.0]
+        for method, expected in [
+            ('inclusive', [0.5, 0.5, 0.8, 1.4]),
+            ('exclusive', [0.5, 0.5, 1.1, 2.3]),
+        ]:
+            with self.subTest(method=method):
+                self.assertEqual(
+                    statistics.quantiles(data, n=5, method=method),
+                    expected,
+                )
 
     def test_equal_sized_groups(self):
         quantiles = statistics.quantiles

Misc/NEWS.d/next/Library/2026-05-23-23-14-26.gh-issue-150318.Utah1I.rst

@@ -1 +1,2 @@
-Fix ``quantiles(method='exclusive')`` returning unsorted cut points for duplicate floats.
+Fix :func:`statistics.quantiles` returning unsorted cut points for duplicate
+floats.