updated_at
自动指数的推荐方法是什么? low查询分析显示日期范围的顺序扫描查询:
created_at
架构定义确认没有索引updated_at
和+----------------------------------------------------------------------------------------------------------------------------------+
| QUERY PLAN |
|----------------------------------------------------------------------------------------------------------------------------------|
| Aggregate (cost=969.33..969.34 rows=1 width=8) (actual time=2.675..2.676 rows=1 loops=1) |
| -> Sort (cost=969.31..969.31 rows=1 width=214) (actual time=2.672..2.672 rows=0 loops=1) |
| Sort Key: analyses.updated_at DESC |
| Sort Method: quicksort Memory: 25kB |
| -> Seq Scan on analyses (cost=0.00..969.30 rows=1 width=214) (actual time=2.666..2.666 rows=0 loops=1) |
| Filter: ((status = 3) AND (updated_at < (now())::date) AND (updated_at >= ((now() + '00:00:10'::interval))::date)) |
| Rows Removed by Filter: 9012 |
| Planning Time: 0.217 ms |
| Execution Time: 2.713 ms
。
created_at
您应该使用以下方式在模型中在模型中索引
updated_at
Indexes:
"analyses_pkey" PRIMARY KEY, btree (id)
"idx_ecosystem_name_version" UNIQUE, btree (ecosystem, name, version)
"idx_analyses_external_id" UNIQUE, btree (external_id)
"idx_analyses_deleted_at" btree (deleted_at)
"idx_analyses_source" btree (source)
字段:
created_at