MySQL tuning

One Basic knowledge of

explain analysis

id	When there are multiple values Big first
select_type
table	Table used
partitions	Partition
type	Connection type （ At least to ref）	system、const、eq_ref、ref...
possible_keys	Possible indexes	Empty means there is no index
key	Index of actual selection	Actually using the index
key_len	Index length	The smaller the better. （ Related to field type and character encoding ）
ref	Which column of the index is referenced
rows	Scanned rows
filtered
extrea

 

Binary tree

Balanced binary trees

B-Tree

B+Tree

B-Tree and B+Tree contrast

• B+Tree The non leaf nodes of the tree only store indexes and do not store data ,B-Tree Tree non leaf nodes store indexes and data .
• B+Tree Leaf node data is indexed . Better for range query data

Two Indexes

Conditions for creating an index

• Fields that are often used as conditions （ Pay attention to the leftmost principle ）
• grouping 、 Sort 、 duplicate removal 、 Uniqueness constraint
• Join fields （ Note that the connection field types are consistent, otherwise the implicit conversion index may be invalid ） As an index

  Index creation is not recommended

• The recorded data in the table is relatively small
• A lot of duplicate data （ Indexing becomes less efficient ）
• Frequently updated fields （ Index maintenance overhead ）

Index failure scenario

• Left blur is used  
• Use OR Some fields of have no index
• String condition not used ‘ ’ Lead up
• Does not conform to the leftmost principle
• Implicit conversion
• Using the expression Or functions （eg: where  age+1=18  .  where substring(name,1,3）='zhang'