案例:改建分区表

现象描述

如下简单SQL语句查询, 性能瓶颈点在normal_date的Scan上。

  1. QUERY PLAN
  2. -----------------------------------------------------------------------------------------------------------------------------------------------
  3. Seq Scan on normal_date (cost=0.00..259.00 rows=30 width=12) (actual time=0.100..3.466 rows=30 loops=1)
  4. Filter: (("time" >= '2022-09-01 00:00:00'::timestamp without time zone) AND ("time" <= '2022-10-01 00:00:00'::timestamp without time zone))
  5. Rows Removed by Filter: 9970
  6. Total runtime: 3.587 ms
  7. (4 rows)

优化分析

从业务层确认表数据(在time字段上)有明显的日期特征,符合分区表的特征。重新规划normal_date表的表定义:字段time为分区键、月为间隔单位定义分区表normal_date_part。修改后结果如下,性能提升近10倍。

  1. QUERY PLAN
  2. -----------------------------------------------------------------------------------------------------------------------------------------------------
  3. Partition Iterator (cost=0.00..480.00 rows=30 width=12) (actual time=0.038..0.085 rows=30 loops=1)
  4. Iterations: 2
  5. -> Partitioned Seq Scan on normal_date_part (cost=0.00..480.00 rows=30 width=12) (actual time=0.049..0.063 rows=30 loops=2)
  6. Filter: (("time" >= '2022-09-01 00:00:00'::timestamp without time zone) AND ("time" <= '2022-10-01 00:00:00'::timestamp without time zone))
  7. Rows Removed by Filter: 31
  8. Selected Partitions: 3..4
  9. Total runtime: 0.360 ms
  10. (7 rows)