MySQL InnoDB 存储引擎 count加速

SELECT COUNT(1) from table USE INDEX (UNIQUE KEY);

比 使用主键索引要快,我测试的是1000W生产环境数据,快了至少3倍.....

EXPLAIN SELECT COUNT(1) from `order` USE INDEX (PRIMARY)\G;
*************************** 1. row ***************************
id: 1
select_type: SIMPLE
table: order
type: index
possible_keys: NULL
key: PRIMARY
key_len: 8
ref: NULL
rows: 1
Extra: Using index
1 row in set (0.01 sec)

EXPLAIN SELECT COUNT(1) from `order` USE INDEX (UNQ_p)\G;
*************************** 1. row ***************************
id: 1
select_type: SIMPLE
table: order
type: index
possible_keys: NULL
key: UNQ_p
key_len: 99
ref: NULL
rows: 1
Extra: Using index
1 row in set (0.00 sec)

这个非聚集索引比主键索引还列宽还长些....
SELECT COUNT(1) from `order` USE INDEX (UNQ_p);
+----------+
| COUNT(1) |
+----------+
| 10984918 |
+----------+
1 row in set (36.60 sec)

SELECT COUNT(1) from `order` USE INDEX (PRIMARY);
+----------+
| COUNT(1) |
+----------+
| 10984918 |
+----------+
1 row in set (1 min 31.57 sec)