MySQL 8 新特性之降序索引

時間 2019-12-05

標籤 mysql 特性降序索引欄目 MySQL 简体版

原文原文鏈接

MySQL 8.0終於支持降序索引了。其實，從語法上，MySQL 4就支持了，但正如官方文檔所言，"they are parsed but ignored"，實際建立的仍是升序索引。

無圖無真相，同一個建表語句，看看MySQL 5.7和8.0的區別。

create table slowtech.t1(c1 int,c2 int,index idx_c1_c2(c1,c2 desc));

MySQL 5.7mysql

mysql> show create table slowtech.t1\G
*************************** 1. row ***************************
      Table: t1
Create Table: CREATE TABLE `t1` (
  `c1` int(11) DEFAULT NULL,
  `c2` int(11) DEFAULT NULL,
  KEY `idx_c1_c2` (`c1`,`c2`)
) ENGINE=InnoDB DEFAULT CHARSET=latin1
1 row in set (0.00 sec)

雖然c2列指定了desc，但在實際的建表語句中仍是將其忽略了。再來看看MySQL 8.0的結果。

mysql> show create table slowtech.t1\G
*************************** 1. row ***************************
      Table: t1
Create Table: CREATE TABLE `t1` (
  `c1` int(11) DEFAULT NULL,
  `c2` int(11) DEFAULT NULL,
  KEY `idx_c1_c2` (`c1`,`c2` DESC)
) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4 COLLATE=utf8mb4_0900_ai_ci
1 row in set (0.00 sec)

c2列仍是保留了desc子句。sql

降序索引的意義

若是一個查詢，須要對多個列進行排序，且順序要求不一致。在這種場景下，要想避免數據庫額外的排序-「filesort」，只能使用降序索引。仍是上面這張表，來看看有降序索引和沒有的區別。

MySQL 5.7

mysql> explain select * from slowtech.t1 order by c1,c2 desc;
+----+-------------+-------+------------+-------+---------------+-----------+---------+------+------+----------+-----------------------------+
| id | select_type | table | partitions | type  | possible_keys | key      | key_len | ref  | rows | filtered | Extra                      |
+----+-------------+-------+------------+-------+---------------+-----------+---------+------+------+----------+-----------------------------+
|  1 | SIMPLE      | t1    | NULL      | index | NULL          | idx_c1_c2 | 10      | NULL |    1 |  100.00 | Using index; Using filesort |
+----+-------------+-------+------------+-------+---------------+-----------+---------+------+------+----------+-----------------------------+
1 row in set, 1 warning (0.00 sec)

MySQL 8.0

mysql> explain select * from slowtech.t1 order by c1,c2 desc;
+----+-------------+-------+------------+-------+---------------+-----------+---------+------+------+----------+-------------+
| id | select_type | table | partitions | type  | possible_keys | key      | key_len | ref  | rows | filtered | Extra      |
+----+-------------+-------+------------+-------+---------------+-----------+---------+------+------+----------+-------------+
|  1 | SIMPLE      | t1    | NULL      | index | NULL          | idx_c1_c2 | 10      | NULL |    1 |  100.00 | Using index |
+----+-------------+-------+------------+-------+---------------+-----------+---------+------+------+----------+-------------+
1 row in set, 1 warning (0.00 sec)

二者的對比能夠看出，MySQL 8.0由於降序索引的存在，避免了「filesort」。

這實際上是降序索引的主要應用場景。若是隻對單個列進行排序，降序索引的意義不是太大，不管是升序仍是降序，升序索引徹底能夠應付。仍是一樣的表，看看下面的查詢。

MySQL 5.7

mysql> explain select * from slowtech.t1 order by c1;
+----+-------------+-------+------------+-------+---------------+-----------+---------+------+------+----------+-------------+
| id | select_type | table | partitions | type  | possible_keys | key      | key_len | ref  | rows | filtered | Extra      |
+----+-------------+-------+------------+-------+---------------+-----------+---------+------+------+----------+-------------+
|  1 | SIMPLE      | t1    | NULL      | index | NULL          | idx_c1_c2 | 10      | NULL |    1 |  100.00 | Using index |
+----+-------------+-------+------------+-------+---------------+-----------+---------+------+------+----------+-------------+
1 row in set, 1 warning (0.00 sec)

mysql> explain select * from slowtech.t1 order by c1 desc;
+----+-------------+-------+------------+-------+---------------+-----------+---------+------+------+----------+-------------+
| id | select_type | table | partitions | type  | possible_keys | key      | key_len | ref  | rows | filtered | Extra      |
+----+-------------+-------+------------+-------+---------------+-----------+---------+------+------+----------+-------------+
|  1 | SIMPLE      | t1    | NULL      | index | NULL          | idx_c1_c2 | 10      | NULL |    1 |  100.00 | Using index |
+----+-------------+-------+------------+-------+---------------+-----------+---------+------+------+----------+-------------+
1 row in set, 1 warning (0.00 sec)

雖然c1是升序索引，但在第二個查詢中，對其進行降序排列時，並無進行額外的排序，使用的仍是索引。在這裏，你們容易產生誤區，覺得升序索引就不能用於降序排列，實際上，對於索引，MySQL不只支持正向掃描，還能夠反向掃描。反向掃描的性能一樣不差。如下是官方對於降序索引的壓測結果，測試表也只有兩列（a，b），建了一個聯合索引(a desc,b asc)，感興趣的童鞋能夠看看，http://mysqlserverteam.com/mysql-8-0-labs-descending-indexes-in-mysql/數據庫

而在8.0中，對於反向掃描，有一個專門的詞進行描述「Backward index scan」。

MySQL 8.0

mysql> explain select * from slowtech.t1 order by c1;
+----+-------------+-------+------------+-------+---------------+-----------+---------+------+------+----------+-------------+
| id | select_type | table | partitions | type  | possible_keys | key      | key_len | ref  | rows | filtered | Extra      |
+----+-------------+-------+------------+-------+---------------+-----------+---------+------+------+----------+-------------+
|  1 | SIMPLE      | t1    | NULL      | index | NULL          | idx_c1_c2 | 10      | NULL |    1 |  100.00 | Using index |
+----+-------------+-------+------------+-------+---------------+-----------+---------+------+------+----------+-------------+
1 row in set, 1 warning (0.00 sec)

mysql> explain select * from slowtech.t1 order by c1 desc;
+----+-------------+-------+------------+-------+---------------+-----------+---------+------+------+----------+----------------------------------+
| id | select_type | table | partitions | type  | possible_keys | key      | key_len | ref  | rows | filtered | Extra                            |
+----+-------------+-------+------------+-------+---------------+-----------+---------+------+------+----------+----------------------------------+
|  1 | SIMPLE      | t1    | NULL      | index | NULL          | idx_c1_c2 | 10      | NULL |    1 |  100.00 | Backward index scan; Using index |
+----+-------------+-------+------------+-------+---------------+-----------+---------+------+------+----------+----------------------------------+
1 row in set, 1 warning (0.00 sec)

終於再也不對group by進行隱式排序sqlserver

因爲降序索引的引入，MySQL 8.0不再會對group by操做進行隱式排序。

下面看看MySQL 5.7和8中的測試狀況

create table slowtech.t1(id int);
insert into slowtech.t1 values(2);
insert into slowtech.t1 values(3);
insert into slowtech.t1 values(1);

MySQL 5.7性能

mysql> select * from slowtech.t1 group by id;
+------+
| id  |
+------+
|    1 |
|    2 |
|    3 |
+------+
3 rows in set (0.00 sec)

mysql> explain select * from slowtech.t1 group by id;
+----+-------------+-------+------------+------+---------------+------+---------+------+------+----------+---------------------------------+
| id | select_type | table | partitions | type | possible_keys | key  | key_len | ref  | rows | filtered | Extra                          |
+----+-------------+-------+------------+------+---------------+------+---------+------+------+----------+---------------------------------+
|  1 | SIMPLE      | t1    | NULL      | ALL  | NULL          | NULL | NULL    | NULL |    3 |  100.00 | Using temporary; Using filesort |
+----+-------------+-------+------------+------+---------------+------+---------+------+------+----------+---------------------------------+
1 row in set, 1 warning (0.00 sec)

「Using filesort」，表明查詢中有排序操做，從結果上看，id列確實也是升序輸出。測試

MySQL 8.0spa

mysql> select * from slowtech.t1 group by id;
+------+
| id  |
+------+
|    2 |
|    3 |
|    1 |
+------+
3 rows in set (0.00 sec)

mysql> explain select * from slowtech.t1 group by id;
+----+-------------+-------+------------+------+---------------+------+---------+------+------+----------+-----------------+
| id | select_type | table | partitions | type | possible_keys | key  | key_len | ref  | rows | filtered | Extra          |
+----+-------------+-------+------------+------+---------------+------+---------+------+------+----------+-----------------+
|  1 | SIMPLE      | t1    | NULL      | ALL  | NULL          | NULL | NULL    | NULL |    3 |  100.00 | Using temporary |
+----+-------------+-------+------------+------+---------------+------+---------+------+------+----------+-----------------+
1 row in set, 1 warning (0.01 sec)