Re: Implementing Incremental View Maintenance
| От | huyajun |
|---|---|
| Тема | Re: Implementing Incremental View Maintenance |
| Дата | |
| Msg-id | tencent_ECCCAEE7BC7D01F3D191DD55DA196B515909@qq.com обсуждение исходный текст |
| Ответ на | Re: Implementing Incremental View Maintenance (Yugo NAGATA <nagata@sraoss.co.jp>) |
| Ответы |
Re: Implementing Incremental View Maintenance
|
| Список | pgsql-hackers |
Hi, Nagata-san
I am very interested in IMMV and read your patch but have some comments in v25-0007-Add-Incremental-View-Maintenance-support.patch and want to discuss with you.
+ /* For IMMV, we need to rewrite matview query */ + query = rewriteQueryForIMMV(query, into->colNames); + query_immv = copyObject(query);
/* Create triggers on incremental maintainable materialized view */ + Assert(query_immv != NULL); + CreateIvmTriggersOnBaseTables(query_immv, matviewOid, true);
1. Do we need copy query?Is it okay that CreateIvmTriggersOnBaseTables directly use (Query *) into->viewQuery instead of query_immv like CreateIndexOnIMMV? It seems only planner may change query, but it shouldn't affect us finding the correct base table in CreateIvmTriggersOnBaseTables .+void
+CreateIndexOnIMMV(Query *query, Relation matviewRel)
+{
+ Query *qry = (Query *) copyObject(query);
2. Also, is it okay to not copy query in CreateIndexOnIMMV? It seems we only read query in CreateIndexOnIMMV.
Regards,
В списке pgsql-hackers по дате отправления: