携手创作,共同成长！这是我参与「掘金日新计划 · 8 月更文挑战」的第2天,点击查看活动详情

一、First look at the index

1、定义

数据库索引,是数据库管理系统(DBMS)中-a sorted data structure,以协助快速查询、更新数据库表中数据.

2、本质

一种有序的数据结构

3、InnoDB中索引类型

索引类型	索引特点
普通（Normal）	最普通的索引,没有任何的限制,也叫非唯一索引.
唯一（Unique）	唯一索引要求键值不能重复.主键索引是特殊的唯一索引（primay key）
全文（Fulltext）	For larger text-type fields（char、verchar、text）进行likeThe query efficiency is slow and a full-text index can be created.

二、MySQL（InnoDB）中的索引结构

1、B+ Tree

Combine the definition and nature of indexes,通俗来说,An index is a data structure that improves query efficiency,而InnoDBThe data structure used isB+ Tree

2、B+ Tree的特性

1、The number of keywords and the number of paths are equal
2、Only store information in leaf nodes,References between keys and leaves are stored on the parent node,Adding further reduces the depth of the tree
3、B+TreeEach leaf node of is added a pointer to the adjacent leaf node,它的最后一个数据会指向下一个叶子节点的第一个数据,形成了一个有序链表的结构.（Improved the speed of traversing requirements）

3、特性带来的优势

1、它是B Tree的升级版,B Tree解决了AVL树(平衡二叉树）A single node has less stored value（The depth of the tree is too deep for large data volumes）的问题（当我们用树的结构来存储索引的时候,Accessing a node happens once with the diskIO操作,InnoDB操作磁盘的最小单位是一页（16k）,Then a tree node must be designed as 16K的大小,Otherwise, there will be situations where the reading is incomplete or not enough.而AVLA node of the tree stores only one key value、地址、子节点引用,远达不到16K,造成浪费.）
2、扫库、扫表能力更强(如果我们要对表进行全表扫描,只需要遍历叶子节点就可以了,不需要遍历整棵B+Tree拿到所有的数据)
3、排序能力更强(因为叶子节点上有下一个数据区的指针,数据形成了链表)
4、效率更加稳定(B+ Tree永远是在叶子节点拿到数据,所以IO次数是稳定的)

4、InnoDB中的B+ Tree

首先了解一个概念：

聚集索引（聚簇索引）：Refers to the index and data clustered together,What is stored on the leaf nodes of the index is not the address value of the usual data in the library,It is directly this data

MyISAM和InnoDB的一个区别就是MyISAMIndex files in the engine（.MYI）和数据文件（.MYD）是分开的,There is no difference in how data is stored and retrieved in a non-primary key index and a primary key index,一样是在索引文件里面找到磁盘地址,然后到数据文件里面获取数据.

而InnoDB使用聚集索引（聚簇索引）的概念,Data is stored directly on the leaves,所以实现索引即数据,数据即索引. 除了聚集索引,Others are called secondary indexes（secondary index）The leaf nodes of the secondary index store the keys of the clustered index,

聚集索引的选取规则：
1、如果我们定义了主键(PRIMARY KEY),那么InnoDB会选择主键作为聚集索引.
2、如果没有显式定义主键,则InnoDB会选择第一个不包含有 NULL值的唯一索引作为主键索引.
3、如果也没有这样的唯一索引, 则InnoDB会选择内置6字节长的ROWID作为隐藏的聚集索引,它会随着行记录的写入而主键递增.

5、索引的使用原则

1、列的离散度

列的重复值越多,离散度越低;重复值越少,离散度就越高.

When we select the index column, we should choose the column with high discreteness as much as possible to build the index.

2、The leftmost match of the column

When creating a joint index, be sure to put the most commonly used columns on the leftmost,This eliminates the need to repeatedly index individual columns.

举个“栗子”：一张user表,列：姓名、性别、身份证号

First, the gender dispersion is too low,不适合建立索引,It is assumed that the ID number is often used as a separate condition for querying in business,When we build a joint index of ID number and name, we need to put the ID number on the left side of the joint index,In this way, you can also use the index when you use the ID number as the query condition alone,No name creation is required、The joint index of ID number and the separate index of ID number have two indexes.

3、覆盖索引

在二级索引里面,不管是单列索引还是联合索引,如果select的数据列只用从索引中就能够取得,不必从数据区中读取,这时候使用的索引就叫做覆盖索引,这样就避免了回表（非主键索引,我们先通过索引找到主键索引的键值,再通过主键值查出索引里面没有的数据,It's better than a bow based on the primary key|The query scans one more index tree, 这个过程就叫回表.）