Populate rooms.creator field for easy lookup

synapse/api/constants.py

@@ -198,6 +198,8 @@ class EventContentFields:
     # cf https://github.com/matrix-org/matrix-doc/pull/1772
     ROOM_TYPE = "type"
 
+    ROOM_CREATOR = "creator"
+
     # Used on normal messages to indicate they were historically imported after the fact
     MSC2716_HISTORICAL = "org.matrix.msc2716.historical"
     # For "insertion" events to indicate what the next chunk ID should be in

synapse/storage/databases/main/room.py

@@ -19,7 +19,7 @@
 from enum import Enum
 from typing import Any, Dict, List, Optional, Tuple
 
-from synapse.api.constants import EventTypes, JoinRules
+from synapse.api.constants import EventContentFields, EventTypes, JoinRules
 from synapse.api.errors import StoreError
 from synapse.api.room_versions import RoomVersion, RoomVersions
 from synapse.events import EventBase
@@ -1012,9 +1012,9 @@ def get_rooms_for_retention_period_in_range_txn(txn):
 class _BackgroundUpdates:
     REMOVE_TOMESTONED_ROOMS_BG_UPDATE = "remove_tombstoned_rooms_from_directory"
     ADD_ROOMS_ROOM_VERSION_COLUMN = "add_rooms_room_version_column"
-    POPULATE_ROOMS_CREATOR_COLUMN = "populate_rooms_creator_column"
     POPULATE_ROOM_DEPTH_MIN_DEPTH2 = "populate_room_depth_min_depth2"
     REPLACE_ROOM_DEPTH_MIN_DEPTH = "replace_room_depth_min_depth"
+    POPULATE_ROOMS_CREATOR_COLUMN = "populate_rooms_creator_column"
 
 
 _REPLACE_ROOM_DEPTH_SQL_COMMANDS = (
@@ -1046,11 +1046,6 @@ def __init__(self, database: DatabasePool, db_conn, hs):
             self._background_add_rooms_room_version_column,
         )
 
-        self.db_pool.updates.register_background_update_handler(
-            _BackgroundUpdates.POPULATE_ROOMS_CREATOR_COLUMN,
-            self._background_populate_rooms_creator_column,
-        )
-
         # BG updates to change the type of room_depth.min_depth
         self.db_pool.updates.register_background_update_handler(
             _BackgroundUpdates.POPULATE_ROOM_DEPTH_MIN_DEPTH2,
@@ -1061,6 +1056,11 @@ def __init__(self, database: DatabasePool, db_conn, hs):
             self._background_replace_room_depth_min_depth,
         )
 
+        self.db_pool.updates.register_background_update_handler(
+            _BackgroundUpdates.POPULATE_ROOMS_CREATOR_COLUMN,
+            self._background_populate_rooms_creator_column,
+        )
+
     async def _background_insert_retention(self, progress, batch_size):
         """Retrieves a list of all rooms within a range and inserts an entry for each of
         them into the room_retention table.
@@ -1198,63 +1198,6 @@ def _background_add_rooms_room_version_column_txn(txn: LoggingTransaction):
 
         return batch_size
 
-    async def _background_populate_rooms_creator_column(
-        self, progress: dict, batch_size: int
-    ):
-        """Background update to go and add creator information to `rooms`
-        table from `current_state_events` table.
-        """
-
-        last_room_id = progress.get("room_id", "")
-
-        def _background_populate_rooms_creator_column_txn(txn: LoggingTransaction):
-            sql = """
-                SELECT room_id, json FROM current_state_events
-                INNER JOIN event_json USING (room_id, event_id)
-                WHERE room_id > ? AND type = 'm.room.create' AND state_key = ''
-                ORDER BY room_id
-                LIMIT ?
-            """
-
-            txn.execute(sql, (last_room_id, batch_size))
-
-            new_last_room_id = ""
-            for room_id, event_json in txn:
-                event_dict = db_to_json(event_json)
-
-                creator = event_dict.get("content").get("creator")
-
-                self.db_pool.simple_update_txn(
-                    txn,
-                    table="rooms",
-                    keyvalues={"room_id": room_id},
-                    updatevalues={"creator": creator},
-                )
-                new_last_room_id = room_id
-
-            if new_last_room_id == "":
-                return True
-
-            self.db_pool.updates._background_update_progress_txn(
-                txn,
-                _BackgroundUpdates.POPULATE_ROOMS_CREATOR_COLUMN,
-                {"room_id": new_last_room_id},
-            )
-
-            return False
-
-        end = await self.db_pool.runInteraction(
-            "_background_populate_rooms_creator_column",
-            _background_populate_rooms_creator_column_txn,
-        )
-
-        if end:
-            await self.db_pool.updates._end_background_update(
-                _BackgroundUpdates.POPULATE_ROOMS_CREATOR_COLUMN
-            )
-
-        return batch_size
-
     async def _remove_tombstoned_rooms_from_directory(
         self, progress, batch_size
     ) -> int:
@@ -1407,6 +1350,65 @@ def process(txn: Cursor) -> None:
 
         return 0
 
+    async def _background_populate_rooms_creator_column(
+        self, progress: dict, batch_size: int
+    ):
+        """Background update to go and add creator information to `rooms`
+        table from `current_state_events` table.
+        """
+
+        last_room_id = progress.get("room_id", "")
+
+        def _background_populate_rooms_creator_column_txn(txn: LoggingTransaction):
+            sql = """
+                SELECT room_id, json FROM event_json
+                INNER JOIN rooms AS room USING (room_id)
+                INNER JOIN current_state_events AS state_event USING (room_id, event_id)
+                WHERE room_id > ? AND (room.creator IS NULL OR room.creator = '') AND state_event.type = 'm.room.create' AND state_event.state_key = ''
+                ORDER BY room_id
+                LIMIT ?
+            """
+
+            txn.execute(sql, (last_room_id, batch_size))
+            room_id_to_create_event_results = txn.fetchall()
+
+            new_last_room_id = ""
+            for room_id, event_json in room_id_to_create_event_results:
+                event_dict = db_to_json(event_json)
+
+                creator = event_dict.get("content").get(EventContentFields.ROOM_CREATOR)
+
+                self.db_pool.simple_update_txn(
+                    txn,
+                    table="rooms",
+                    keyvalues={"room_id": room_id},
+                    updatevalues={"creator": creator},
+                )
+                new_last_room_id = room_id
+
+            if new_last_room_id == "":
+                return True
+
+            self.db_pool.updates._background_update_progress_txn(
+                txn,
+                _BackgroundUpdates.POPULATE_ROOMS_CREATOR_COLUMN,
+                {"room_id": new_last_room_id},
+            )
+
+            return False
+
+        end = await self.db_pool.runInteraction(
+            "_background_populate_rooms_creator_column",
+            _background_populate_rooms_creator_column_txn,
+        )
+
+        if end:
+            await self.db_pool.updates._end_background_update(
+                _BackgroundUpdates.POPULATE_ROOMS_CREATOR_COLUMN
+            )
+
+        return batch_size
+
 
 class RoomStore(RoomBackgroundUpdateStore, RoomWorkerStore, SearchStore):
     def __init__(self, database: DatabasePool, db_conn, hs):
@@ -1438,9 +1440,9 @@ async def upsert_room_on_join(
             # invalid, and it would fail auth checks anyway.
             raise StoreError(400, "No create event in state")
 
-        room_creator = create_event.content.get("creator", None)
+        room_creator = create_event.content.get(EventContentFields.ROOM_CREATOR)
 
-        if room_creator is None:
+        if not isinstance(room_creator, str):
             # If the create event does not have a creator then the room is
             # invalid, and it would fail auth checks anyway.
             raise StoreError(400, "No creator defined on the create event")

synapse/storage/schema/main/delta/63/02populate-rooms-creator.sql

@@ -13,5 +13,5 @@
  * limitations under the License.
  */
 
-INSERT into background_updates (update_name, progress_json)
+INSERT INTO background_updates (update_name, progress_json)
     VALUES ('populate_rooms_creator_column', '{}');