What are the knowledge points about MySQL indexing and optimization?

What is index?

• Index is a data structure that helps MySQL perform efficient queries. Like the table of contents of a book, it can speed up the query

The structure of the index?

The index can have B-Tree index and Hash index. The index is implemented in the storage engine

InnoDB/MyISAM only supports B-Tree index

Memory/Heap supports B-Tree index and Hash index

• B-Tree

  B-Tree is a data structure that is very suitable for disk operations. It is a multi-way balanced search tree. Its height is generally 2-4, and its non-leaf nodes and leaf nodes will store data. All its leaf nodes are on the same layer. The picture below is a B-Tree

• ## B Tree: B-Tree is a type based on B-Tree optimization. The main difference between it and the B-tree is that all the data of the B-tree is stored in the leaf nodes, and the leaf nodes are strung together by a linked list. The following picture is a B-tree

The size of a page in InnoDB is 16KB (a page is a node on the B-tree). If the table The primary key is INT and the size is 4 bytes. That node can also store 4K key values. Assuming that both pointers and key values ​​occupy the same size, then the B-tree with a height of 3 has 2048 nodes in the second layer and 2048 nodes in the third layer. The number of leaf nodes in the layer is 2048*2048 = 4194304, and one node is 16KB, so a total of 67108864KB, or 65536MB, or 64G of data can be accommodated.

Since the leaf nodes are strung together by a linked list, if ordered by the index column, it will be sorted by default, so the efficiency will be very high.

• MyISAM index

  MyISAM index and data are stored separately. In the primary key index of MyISAM, the record address is stored in the leaf node of the B tree, so MyISAM needs to go through 2 IOs through the index query
  

MyISAM's auxiliary index is the same as the primary key index. The only difference is that the key in the auxiliary index can be repeated, but the key in the primary key index cannot be repeated.

• InnoDB index

  InnoDB Data and indexes are stored together, also called clustered indexes. Data is indexed by the primary key and stored on the leaf nodes of the primary key index B-tree.
  InnoDB primary key index, the data is already included in the leaf nodes, that is, the index and data are stored together, which is a clustered index.
  

In the auxiliary index of InnoDB, the primary key value is stored in the leaf node, not the address. Using the auxiliary index requires two searches.

The difference between InnoDB and MyISAM indexes:

• InnoDB uses a clustered index, and its primary key index directly stores data in the leaf nodes. The leaf nodes in its auxiliary index store the value of the primary key

• MyISAM uses a non-clustered index. The data and index are not in the same file. The leaf nodes in its primary key index store The address where the row record is located, and the leaf node in the auxiliary index also stores the address where the record is located, but the key of the auxiliary index can be repeated, but the key of the primary key index cannot be repeated

  
  

Question:

• Why doesn’t InnoDB use overly long fields as primary keys? A primary key that is too long will make the auxiliary index take up a lot of space
  

• Why is it recommended that InnoDB use auto-incrementing primary keys? If you use an auto-incrementing primary key, each time a new record is inserted, the new record will be sequentially added to the subsequent position of the current index node. When one page is full, a new page will be opened, so that The index structure is very compact, and there is no need to move existing data every time it is inserted, which is very efficient. If you do not use an auto-incrementing primary key, each time you insert a new record, you have to select an insertion position, and the data may need to be moved, which makes the efficiency not high and the index structure not compact
  

• Why use B-tree instead of B-tree

Where does the index exist?

• The index itself is also relatively large and is generally stored on the disk. The index and data may be stored separately (MyISAM's non-clustered index) or they may be stored together (InnoDB's Clustered index)

What are the advantages and disadvantages of index?

• Advantages

• Reduce IO Cost, improve data query efficiency

• Reduce sorting cost (the indexed columns will be automatically sorted, using order by will improve the efficiency a lot)

• shortcoming

• The index will occupy additional storage space

• The index will reduce the efficiency of updating table data. When adding, deleting, or modifying operations, you must not only save the data, but also update the corresponding index

Classification of the index

• Single column index

• Primary key index

• Unique index

• Normal index

• Combined index

Index uses

• to create an index

 CREATE INDEX index_name ON table_name(col_name);
-- 或者
ALTER TABLE table_name ADD INDEX index_name(col_name)

• Delete index

DROP INDEX index_name ON table_name;

• Scenarios that require indexing

• Frequent The columns used as query conditions need to be indexed

• In multi-table association, the associated fields need to be indexed

• The sorting fields in the query need to be indexed Building an index

• Scenarios where indexing is not suitable

• ##A table that writes more and reads less is not suitable for indexing

• Frequently updated fields are not suitable for building indexes

explain execution plan

There is an existing user table, in which The index is as follows

The three fields of name, age and address are used as a combined index

You can use explain to perform performance analysis on a certain SQL statement

explain select * from user where name = 'am';

possible_keysPossibly used index

keyActually used index

key_lenThe length of the index used for query

refIf it is an equivalent query, it will be const

rowsEstimated number of rows to be scanned (not an exact value)

extra

Extra information, such as

• using where

  indicates that the results returned by the storage engine also need to be filtered at the SQL Layer layer
  

• using index

  indicates that there is no need to query back to the table. Generally, it will be when using a covering index. this value. Covering index means that the columns in the select are all index columns. The query that does not need to be returned to the table means that you can get the value of the index column directly by going through the auxiliary index, and there is no need to go to the primary key index to fetch records
  

• using index condition

  MySQL 5.6.x and later supports the ICP feature (Index Condition Pushdown), which can push the check conditions to the storage engine layer. Records that do not meet the conditions will not be read directly, instead of reading them out first and then reading them as before. SQL Layer layer filtering, which reduces the number of rows scanned by the storage engine layer
  

##using filesort
• cannot be sorted Index used

  
  

type

system: There is only 1 row of data in the table, or the table is empty
const: Use a unique index or primary key index, and query with where and other values. The returned record is 1 row, also called a unique index scan

#ref: For non-unique indexes, queries using equivalent where conditions or leftmost prefix rules.
The following is the leftmost prefix rule that is satisfied, that is, for idx_name_age_add, the leftmost prefix is ​​satisfied, and the first index is name

range: Index range scan, common in >, <, between, in, like and other queries

##Note that when like, the wildcard character % cannot be placed at the beginning, otherwise it will cause a full table scan

##index ：

#all that does not completely match the index, but does not need to be queried back to the table : Scan the entire table, and then filter the records that meet the requirements in the SQL Layer

索引使用规范（索引失效分析）

1. 全值匹配
   在索引列上使用等值查询

explain select * from user where name = 'y' and age = 15;

2. 最左前缀

组合索引中，查询条件要从组合索引的最左列开始，如上述example中组合索引idx_name_age_add，是建立在三个列name，age，address的，若跳过name，直接用age查询，则会变为全表扫描

explain select * from user where age = 15;

3. 不要在索引列上做计算

4. 范围条件右侧的索引列会失效

看到第一个SQL语句，没有用上addresss索引

5. 尽量使用覆盖索引

explain select name,age from user where name = 'y' and age = 1;

可以避免回表查询

6. 索引字段不要使用不等（!= 或 ）,不要判断null（is null/ is not null）
会导致索引失效，转为全表扫描

7. 索引字段上使用like时，不要以%开头

8. 索引字段如果是字符串，记得加单引号

9. 索引字段不要用or

例子总结：

The above is the detailed content of What are the knowledge points about MySQL indexing and optimization?. For more information, please follow other related articles on the PHP Chinese website!