Preserve NDV precision in max_distinct_count when cap is not triggered

Author: asolimando

datafusion/physical-plan/src/joins/utils.rs

@@ -712,7 +712,13 @@ fn max_distinct_count(
             // NDV can never exceed the number of rows
             match num_rows {
                 Precision::Absent => dc,
-                _ => dc.min(num_rows).to_inexact(),
+                _ => {
+                    if dc.get_value() <= num_rows.get_value() {
+                        dc
+                    } else {
+                        num_rows.to_inexact()
+                    }
+                }
             }
         }
         _ => {
@@ -2974,4 +2980,76 @@ mod tests {
         let result = max_distinct_count(&num_rows, &stats);
         assert_eq!(result, Exact(0));
     }
+
+    #[test]
+    fn test_max_distinct_count_preserves_precision_when_not_capped() {
+        assert_eq!(
+            max_distinct_count(
+                &Exact(10),
+                &ColumnStatistics {
+                    distinct_count: Exact(5),
+                    ..Default::default()
+                }
+            ),
+            Exact(5)
+        );
+        assert_eq!(
+            max_distinct_count(
+                &Exact(10),
+                &ColumnStatistics {
+                    distinct_count: Inexact(5),
+                    ..Default::default()
+                }
+            ),
+            Inexact(5)
+        );
+        // Inexact num_rows does not affect an exact NDV that is within bounds
+        assert_eq!(
+            max_distinct_count(
+                &Inexact(10),
+                &ColumnStatistics {
+                    distinct_count: Exact(5),
+                    ..Default::default()
+                }
+            ),
+            Exact(5)
+        );
+    }
+
+    #[test]
+    fn test_max_distinct_count_demotes_to_inexact_when_capped() {
+        // Exact NDV > Exact num_rows is an illegal state (NDV <= num_rows is a
+        // mathematical invariant), but the code handles it defensively by
+        // capping and demoting to inexact
+        assert_eq!(
+            max_distinct_count(
+                &Exact(10),
+                &ColumnStatistics {
+                    distinct_count: Exact(15),
+                    ..Default::default()
+                }
+            ),
+            Inexact(10)
+        );
+        assert_eq!(
+            max_distinct_count(
+                &Inexact(10),
+                &ColumnStatistics {
+                    distinct_count: Exact(15),
+                    ..Default::default()
+                }
+            ),
+            Inexact(10)
+        );
+        assert_eq!(
+            max_distinct_count(
+                &Exact(10),
+                &ColumnStatistics {
+                    distinct_count: Inexact(15),
+                    ..Default::default()
+                }
+            ),
+            Inexact(10)
+        );
+    }
 }