Fix deterministic fingerprints for set-valued arguments

src/databricks/labs/dqx/utils.py

@@ -189,8 +189,8 @@ def normalize_bound_args(val: Any, allow_simple_expressions_only: bool = True) -
     """
     Normalize a value or collection of values for consistent processing.
 
-    Handles primitives, dates, Decimal, and column-like objects. Lists, tuples, and sets are
-    recursively normalized with type preserved.
+    Handles primitives, dates, Decimal, and column-like objects. Collections are recursively
+    normalized, preserving list and tuple order while canonicalizing set and frozenset order.
 
     For Decimal values, uses a special JSON-serializable format to preserve type information
     for round-trip deserialization.
@@ -211,7 +211,11 @@ def normalize_bound_args(val: Any, allow_simple_expressions_only: bool = True) -
     if val is None:
         return None
 
-    if isinstance(val, (list, tuple, set, frozenset)):
+    if isinstance(val, (set, frozenset)):
+        normalized = [normalize_bound_args(v, allow_simple_expressions_only) for v in val]
+        return sorted(normalized, key=lambda item: json.dumps(item, sort_keys=True, default=str))
+
+    if isinstance(val, (list, tuple)):
         return [normalize_bound_args(v, allow_simple_expressions_only) for v in val]
 
     if isinstance(val, dict):

tests/unit/test_rule.py

@@ -1,5 +1,7 @@
 """Unit tests for rule fingerprinting, replacement, expansion, and serialization alignment."""
 
+from collections.abc import Iterator
+from enum import Enum
 import re
 
 import pytest
@@ -31,6 +33,29 @@ def test_compute_rule_fingerprint_same_rule_same_fingerprint():
     assert _hex_sha256_pattern().match(fp1)
 
 
+def test_compute_rule_fingerprint_set_arguments_order_independent():
+    """Set arguments with mixed nested values produce the same fingerprint in any iteration order."""
+
+    class NestedValue(Enum):
+        DICT = {"nested": [1, "1"]}
+
+    class IterationOrderedSet(set[object]):
+        def __init__(self, values: list[object]) -> None:
+            super().__init__(values)
+            self.iteration_order = values
+
+        def __iter__(self) -> Iterator[object]:
+            return iter(self.iteration_order)
+
+    values: list[object] = [NestedValue.DICT, 1, "1", ("pair", 2), frozenset({3, "3"})]
+    check = {"check": {"function": "is_in_list", "arguments": {"allowed": IterationOrderedSet(values)}}}
+    reversed_check = {
+        "check": {"function": "is_in_list", "arguments": {"allowed": IterationOrderedSet(list(reversed(values)))}}
+    }
+
+    assert compute_rule_fingerprint(check) == compute_rule_fingerprint(reversed_check)
+
+
 def test_compute_rule_set_fingerprint_by_metadata_same_set_same_fingerprint():
     """Same rule set produces the same rule_set_fingerprint (determinism)."""
     checks = [