前言
作为DBA了解InnoDB的页组织方式是最基础的,在实际工作中,免不了会评估SQL会消耗多少IO,怎么评估呢?
作为InnoDB表和树的高度或者深度有关系。
查看树的高度?
之前研究了半天:
https://www.percona.com/blog/2009/04/28/the_depth_of_a_b_tree/
http://code.openark.org/blog/mysql/the-depth-of-an-index-primer
根据
Scholmi notes that there are two main features determining the depth of a B-tree (or B+-tree):
The number of rows in the database. We’ll call that N.
The size of the indexed key. Let’s call B the number of key that fit in a B-tree node. (Sometimes B is used to refer to the node size itself, rather than the number of keys it holds, but I hope my choice will make sense directly.)
Given these quantities, the depth of a B-tree is logB N, give or take a little. That’s just (log N)/log B. Now we can rephrase Scholmi’s point as noting that small keys means a bigger B, which reduces (log N)/log B. If we cut the key size in half, then the depth of the B-tree goes from (log N)/log B to (log N)/log 2B (twice as many keys fit in the tree nodes), and that’s just (log N)/(1+log B).
Let’s put some numbers in there. Say you have a billion rows, and you can currently fit 64 keys in a node. Then the depth of the tree is (log 109)/ log 64 ≈ 30/6 = 5. Now you rebuild the tree with keys half the size and you get log 109 / log 128 ≈ 30/7 = 4.3. Assuming the top 3 levels of the tree are in memory, then you go from 2 disk seeks on average to 1.3 disk seeks on average, for a 35% speedup.
里面算的不对,人肉算也没达到这个高度。也可能是我没有理解作者的意思,没有用对公式。那么根据结合前面的Innodb页结构,如何正确的获取数的高度呢?继续拿这个表举例子:
mysql> show create table sbtest1\G
*************************** 1. row ***************************
Table: sbtest1
Create Table: CREATE TABLE `sbtest1` (
`id` int(11) NOT NULL AUTO_INCREMENT,
`gmt_create` datetime NOT NULL,
`gmt_modified` datetime NOT NULL,
`k` int(11) NOT NULL DEFAULT '0',
`c` varchar(500) NOT NULL DEFAULT '',
`pad` char(60) NOT NULL DEFAULT '',
`is_used` int(11) DEFAULT NULL,
PRIMARY KEY (`id`),
KEY `k_1` (`k`),
KEY `idx_is_used` (`is_used`),
KEY `idx_gmt_create` (`gmt_create`)
) ENGINE=InnoDB AUTO_INCREMENT=69313841 DEFAULT CHARSET=latin1
1 row in set (0.00 sec)
方法一
[root@localhost mysql]# innodb_space -f test/sbtest1.ibd space-index-pages-summary | head -n 10
page index level data free records
3 74 2 5166 10904 369
4 75 2 408 15834 24
5 76 2 486 15756 27
6 77 2 486 15756 27
7 74 0 15028 1192 68
8 74 0 15028 1192 68
9 74 1 14700 1030 1050
10 74 0 15028 1192 68
11 74 0 15028 1192 68
page_level是2,所以这个树高度是page_level+1=3;
方法二
mysql> show global variables like "%innodb_page%";
+----------------------+-------+
| Variable_name | Value |
+----------------------+-------+
| innodb_page_cleaners | 1 |
| innodb_page_size | 16384 |
+----------------------+-------+
2 rows in set (0.00 sec)
mysql> show table status like 'sbtest1'\G
*************************** 1. row ***************************
Name: sbtest1
Engine: InnoDB
Version: 10
Row_format: Dynamic
Rows: 25926320
Avg_row_length: 279
Data_length: 7254032384
Max_data_length: 0
Index_length: 1293697024
Data_free: 3145728
Auto_increment: 69313841
Create_time: 2018-01-19 14:53:11
Update_time: NULL
Check_time: NULL
Collation: latin1_swedish_ci
Checksum: NULL
Create_options:
Comment:
1 row in set (0.00 sec)
mysql> desc sbtest1;
+--------------+--------------+------+-----+---------+----------------+
| Field | Type | Null | Key | Default | Extra |
+--------------+--------------+------+-----+---------+----------------+
| id | int(11) | NO | PRI | NULL | auto_increment |
| gmt_create | datetime | NO | MUL | NULL | |
| gmt_modified | datetime | NO | | NULL | |
| k | int(11) | NO | MUL | 0 | |
| c | varchar(500) | NO | | | |
| pad | char(60) | NO | | | |
| is_used | int(11) | YES | MUL | NULL | |
+--------------+--------------+------+-----+---------+----------------+
通常一棵B+树可以存放多少行数据?
这里我们先假设B+树高为2,即存在一个根节点和若干个叶子节点,那么这棵B+树的存放总记录数为:根节点指针数*单个叶子节点记录行数。
上文我们已经说明单个叶子节点(页)中的记录数=16K/279=58。(我们从上面可以看到每行记录的数据平均大小为279个字节)。
那么现在我们需要计算出非叶子节点能存放多少指针,其实这也很好算,表中的主键ID为int类型,长度为4字节,而指针大小在InnoDB源码中设置为6字节,这样一共10字节,我们一个页中能存放多少这样的单元,其实就代表有多少指针,即16384/10=1638。那么可以算出一棵高度为2的B+树,能存放1638*58=95004条这样的数据记录。
根据同样的原理我们可以算出一个高度为3的B+树可以存放:1638*1638*58=155616552条这样的记录。
高度为4的B+树可以存放:
1638*1638*1638*58=254899912176条这样的记录。
而在实际应用中,大部分是以bigint作为主键的,主键ID为bigint类型,长度为8字节,而指针大小在InnoDB源码中设置为6字节,这样一共14字节,我们一个页中能存放多少这样的单元,其实就代表有多少指针,即16384/14=1170。
根据同样的原理我们可以算出一个高度为2,3,4的B+树能够存放的记录数。
mysql> select count(*) from sbtest1;
+----------+
| count(*) |
+----------+
| 26313272 |
+----------+
1 row in set (4.23 sec)
我们的表一共是3层。
方法三
mysql> SELECT b.name, a.name, index_id, type, a.space, a.PAGE_NO FROM information_schema.INNODB_SYS_INDEXES a, information_schema.INNODB_SYS_TABLES b WHERE a.table_id = b.table_id AND a.space <> 0 and b.name='test/sbtest1';
+--------------+----------------+----------+------+-------+---------+
| name | name | index_id | type | space | PAGE_NO |
+--------------+----------------+----------+------+-------+---------+
| test/sbtest1 | PRIMARY | 74 | 3 | 45 | 3 |
| test/sbtest1 | k_1 | 75 | 0 | 45 | 4 |
| test/sbtest1 | idx_is_used | 76 | 0 | 45 | 5 |
| test/sbtest1 | idx_gmt_create | 77 | 0 | 45 | 6 |
+--------------+----------------+----------+------+-------+---------+
4 rows in set (0.00 sec)
primary key的高度是3,其他索引的可以看上表。
方法四
因为主键索引B+树的根页在整个表空间文件中的第3个页开始,所以可以算出它在文件中的偏移量:16384*3=49152(16384为页大小)。
另外根据《InnoDB存储引擎》中描述在根页的64偏移量位置前2个字节,保存了page level的值,因此我们想要的page level的值在整个文件中的偏移量为:16384*3+64=49152+64=49216,前2个字节中。
接下来我们用hexdump工具,查看表空间文件指定偏移量上的数据:
[root@localhost test]# hexdump -s 49216 -n 10 sbtest1.ibd
000c040 0200 0000 0000 0000 4a00
000c04a
page_level是2,B+树高度为page level+1=3
如果通过二级索引查找记录最多需要花费多少次IO呢?
从上面的图中可以看出需要花费:
从二级索引找到主键+主键找到记录,比如二级索引有3层,聚簇索引有3层,那么最多花费的IO次数是:3+3=6
参考
姜承尧 《MySQL技术内幕:InnoDB存储引擎》
姜承尧 http://www.innomysql.com/查看-innodb表中每个的索引高度/
https://blog.jcole.us/2013/01/10/btree-index-structures-in-innodb/?spm=a2c4e.11153940.0.0.77d994fcZH1hIv