fix(idempotency): sorting keys before hashing

Author: heitorlessa

Diff for: aws_lambda_powertools/utilities/idempotency/persistence/base.py

@@ -223,7 +223,7 @@ def _generate_hash(self, data: Any) -> str:
 
         """
         data = getattr(data, "raw_event", data)  # could be a data class depending on decorator order
-        hashed_data = self.hash_function(json.dumps(data, cls=Encoder).encode())
+        hashed_data = self.hash_function(json.dumps(data, cls=Encoder, sort_keys=True).encode())
         return hashed_data.hexdigest()
 
     def _validate_payload(self, data: Dict[str, Any], data_record: DataRecord) -> None:
@@ -310,7 +310,7 @@ def save_success(self, data: Dict[str, Any], result: dict) -> None:
         result: dict
             The response from function
         """
-        response_data = json.dumps(result, cls=Encoder)
+        response_data = json.dumps(result, cls=Encoder, sort_keys=True)
 
         data_record = DataRecord(
             idempotency_key=self._get_hashed_idempotency_key(data=data),

Diff for: tests/functional/idempotency/test_idempotency.py

@@ -976,3 +976,20 @@ def lambda_handler(event, _):
     # THEN we expect the function and lambda handler to execute successfully
     assert fn_result == expected_result
     assert handler_result == expected_result
+
+
+def test_idempotent_data_sorting():
+    # Scenario to validate same data in different order hashes to the same idempotency key
+    data_one = {"data": "test message 1", "more_data": "more data 1"}
+    data_two = {"more_data": "more data 1", "data": "test message 1"}
+
+    # Assertion will happen in MockPersistenceLayer
+    persistence_layer = MockPersistenceLayer("test-func#" + hashlib.md5(json.dumps(data_one).encode()).hexdigest())
+
+    # GIVEN
+    @idempotent_function(data_keyword_argument="payload", persistence_store=persistence_layer)
+    def dummy(payload):
+        return {"message": "hello"}
+
+    # WHEN
+    dummy(payload=data_two)