PostgreSQL 中 index scan 与 seq scan 的对比学习

数据量很小的时候，我们可以看到，seq scan 比 index scan 更加有效。那是因为 index scan 至少要发生两次I/O，一次是 读取索引块， 一次是读取数据块。当index 很大的时候，情况可能会更加复杂。

postgres=# select a.relpages,a.reltuples,a.relfilenode,a.reltype,b.typname from pg_class a,pg_type b where a.relname like 'gaotab%' and a.reltype=b.oid;

relpages | reltuples | relfilenode | reltype | typname ----------+-----------+-------------+---------+---------

1   | 100   | 16396   | 16386 | gaotab

数据量为 100条记录。

预估成本：

set session enable_seqscan=false;

SET

postgres=# explain select name from gaotab where id=50;

QUERY PLAN ---------------------------------------------------------------------

Index Scan using idx_id on gaotab (cost=0.00..8.27 rows=1 width=5)

 Index Cond: (id = 50) (2 rows)

true;

QUERY PLAN ------------------------------------------------------

Seq Scan on gaotab (cost=0.00..2.25 rows=1 width= Filter: (id = 2 rows)

实际执行：

postgres=# explain analyze QUERY PLAN -------------------------------------------------------------------------------- -------------------------------

5) (actual tim e=0.112..0.113 rows=1 loops=1)

50) Total runtime: 0.133 ms (3 rows)

QUERY PLAN -------------------------------------------------------------------------------- ----------------

Seq Scan on gaotab (cost=2.25 rows=5) (actual time=0.014..0.018 rows= Filter: (id = 50) Rows Removed by Filter: 99 Total runtime: 0.034 ms (4 rows)

等到数据量大的时候，就是截然不同了。

数据为1000条记录时，通过查询可以看到，已经跨越了7个page：

postgres=# analyze;

ANALYZE

postgres=# 7   | 1000   | 16386 | gaotab (1 row)

postgres=#

再次预估成本，此时seq scan 已经开始变得不划算了：

postgres=# explain select name from gaotab where id=QUERY PLAN ---------------------------------------------------------------------

using idx_id on gaotab (cost=8.276)

 Index Cond: (id 2 rows)

=# set session enable_seqscan= session enable_indexscan=Bitmap Heap Scan on gaotab (cost4.26 Recheck Cond: (id 50)

 -> Bitmap Index Scan on idx_id (cost=0)

4 rows)

 session enable_bitmapscan=-------------------------------------------------------

Seq Scan on gaotab (cost19.50=#

实际执行

QUERY PLAN ------------------------------------------------------------------------------------------------------------

10000000000.00..10000000008.27 rows=6) (actual time=0.020. .0.022 rows=1)  Index Cond: (id = 0.051 ms (3 rows)

set session enable_indexscan=set session enable_bitmapscan=QUERY PLAN -------------------------------------------------------------------------------------------------

19.50 rows=0.015..0.095 rows=999 Total runtime: 0.109 ms (postgres=#