diff --git a/tap_postgres/stream_utils.py b/tap_postgres/stream_utils.py
index 84a4f46e..aed15824 100644
--- a/tap_postgres/stream_utils.py
+++ b/tap_postgres/stream_utils.py
@@ -72,27 +72,27 @@ def refresh_streams_schema(conn_config: Dict, streams: List[Dict]):
             for stream in discover_db(conn, conn_config.get('filter_schemas'), [st['table_name'] for st in streams])
         }
 
-        LOGGER.debug('New discovery schemas %s', new_discovery)
-
-        # For every stream dictionary, update the schema and metadata from the new discovery
-        for idx, stream in enumerate(streams):
-            # update schema
-            streams[idx]['schema'] = copy.deepcopy(new_discovery[stream['tap_stream_id']]['schema'])
-
-            # Update metadata
-            #
-            # 1st step: new discovery doesn't contain non-discoverable metadata: e.g replication method & key, selected
-            # so let's copy those from the original stream object
-            md_map = metadata.to_map(stream['metadata'])
-            meta = md_map.get(())
-
-            for idx_met, metadatum in enumerate(new_discovery[stream['tap_stream_id']]['metadata']):
-                if not metadatum['breadcrumb']:
-                    meta.update(new_discovery[stream['tap_stream_id']]['metadata'][idx_met]['metadata'])
-                    new_discovery[stream['tap_stream_id']]['metadata'][idx_met]['metadata'] = meta
-
-            # 2nd step: now copy all the metadata from the updated new discovery to the original stream
-            streams[idx]['metadata'] = copy.deepcopy(new_discovery[stream['tap_stream_id']]['metadata'])
+    LOGGER.debug('New discovery schemas %s', new_discovery)
+
+    # For every stream, update the schema and metadata from the new discovery
+    for idx, stream in enumerate(streams):
+        # Update schema
+        streams[idx]['schema'] = copy.deepcopy(new_discovery[stream['tap_stream_id']]['schema'])
+
+        # Create updated metadata
+        original_stream_metadata_map = metadata.to_map(stream['metadata'])
+        new_discovery_metadata_map = metadata.to_map(new_discovery[stream['tap_stream_id']]['metadata'])
+
+        for metadata_element_key in new_discovery_metadata_map:
+            if metadata_element_key in original_stream_metadata_map:
+                original_stream_metadata_map[metadata_element_key].update(new_discovery_metadata_map[metadata_element_key])
+            else:
+                original_stream_metadata_map[metadata_element_key] = new_discovery_metadata_map[metadata_element_key]
+
+        updated_original_metadata_list = metadata.to_list(original_stream_metadata_map)
+
+        # Copy the updated metadata back into the original data structure that was passed in
+        streams[idx]['metadata'] = copy.deepcopy(updated_original_metadata_list)
 
     LOGGER.debug('Updated streams schemas %s', streams)
 
diff --git a/tests/test_full_table_interruption.py b/tests/test_full_table_interruption.py
index 21bc0144..14b7cfca 100644
--- a/tests/test_full_table_interruption.py
+++ b/tests/test_full_table_interruption.py
@@ -99,7 +99,7 @@ def test_catalog(self):
 
         self.assertTrue(blew_up_on_cow)
 
-        self.assertEqual(7, len(CAUGHT_MESSAGES))
+        self.assertEqual(7, len(CAUGHT_MESSAGES), "Number of Caught Messages")
 
         self.assertEqual(CAUGHT_MESSAGES[0]['type'], 'SCHEMA')
         self.assertIsInstance(CAUGHT_MESSAGES[1], singer.StateMessage)
diff --git a/tests/test_streams_utils.py b/tests/test_streams_utils.py
index 114ab5c2..0734c7e5 100644
--- a/tests/test_streams_utils.py
+++ b/tests/test_streams_utils.py
@@ -51,7 +51,14 @@ def test_refresh_streams_schema(self):
 							'table-key-properties': ['some_id'],
 							'row-count': 1000,
 						}
+					},
+					{
+						'breadcrumb': ['properties', 'char_name'],
+						'metadata': {
+							'arbitrary_field_metadata': 'should be preserved'
+						}
 					}
+
 				]
 			}
 		]
@@ -86,7 +93,8 @@ def test_refresh_streams_schema(self):
 										  'selected-by-default': True},
 			('properties', 'char_name'): {'selected-by-default': True,
 										  'inclusion': 'available',
-										  'sql-datatype': 'character'}})
+										  'sql-datatype': 'character',
+										  'arbitrary_field_metadata': 'should be preserved'}})
 
 		self.assertEqual({'properties': {'id': {'type': ['integer'],
 												'maximum': 2147483647,