MySQL indexes and transactions

 

---

Catalog

🥬 What is the index

🥬 Index usage scenarios

🥬 Common operations of indexing

🥬 Data structure behind index

🥬 What is the business

🥬 The basic characteristics of a transaction

🥬 Summary

---

🥬 What is the index

      Index is a special kind of file , Contains a reference pointer to all records in the data table . You can index one or more columns in a table , And specify the type of index , Each index has its own data structure implementation . The index is equivalent to the catalogue of a Book , Through the index Fast positioning 、 Retrieving data . Using index greatly improves the efficiency of search , But at the same time, the index also takes up more space , Slow down the speed of addition, deletion and modification .

🥬 Index usage scenarios

1、 Large amount of data , And we often make conditional queries on these columns .
2、 The insert operation of the database table , And modify these columns less frequently .
3、 Indexes take up extra disk space .
When the above conditions are met , Consider indexing these fields in the table , To improve query efficiency .
conversely , If the non conditional query column , Or insert it often 、 Modify the operating , Or when there is not enough disk space , Don't think about creating indexes .

🥬 Common operations of indexing

1、 Create index

For non primary keys 、 Non unique constraint 、 Non foreign key fields , You can create a normal index .

create index  Index name  on  Table name ( Field name );

2、 Look at the index

show index from  Table name ;

3、 Delete index

drop index  Index name  on  Table name ;

🥬 Data structure behind index

The data structure behind the index is B+ Trees .

 B+ Tree advantage ：

1、 Use B+ When searching the tree, the whole disk IO The number of times is relatively small .

2、 All queries will eventually fall on the leaf node , Disk per query IO The times are the same , The query speed is relatively stable .

3、 After the leaf nodes are connected with a linked list , Very suitable for range finding .

4、 All data storage ( load ) They are all placed on leaf nodes , Only... Are saved in non leaf nodes key value , Therefore, the overall space occupied by non leaf nodes is small , It can even be cached in memory （ Once you can put it all in memory , At this time, the disk IO It's almost gone ）.

🥬 What is the business

      A transaction is a logical set of operations , The units that make up this set of operations , All or nothing , All or nothing .

In different environments , You can have business . In the database , It's database transactions . That is to say, either both operations are performed , Or none of them .

     For example, there are A、B Two accounts , Both accounts have 1000 element , Now? A To give B turn 500 element , that A There should be less in the account 500 element ,B More accounts 500, Both operations should be performed . There cannot be one execution and one non execution , Either not .

     stay sql in , If the first one sql After performing , the second sql Unexpected while executing , Result in failure to execute , At this time, the database will be restored , Let the first sql Restore to before execution （ Roll back ）. So how is the database restored ？ Because the database will record every operation performed .

Transaction related operations

 

1、 Open transaction

start transaction;

2、 Roll back or commit

rollback/commit;
//rollback It's all failure ,commit It's all success .

🥬 The basic characteristics of a transaction

1、 Atomicity

Atomicity means that either both operations are performed , Or none of them , Operations cannot be subdivided .

2、 Uniformity

Consistency means that the data in the database should be reasonable and legal before and after the transaction execution .（ For example, the above transfer , The account cannot be negative ）

3、 persistence

Persistence means that once a transaction is committed , The data is written to the hard disk , Persistent storage .

4、 Isolation,

Isolation refers to the situation caused by the concurrent execution of transactions .

Concurrent transaction execution refers to the simultaneous execution of multiple transactions , When multiple transactions are executed concurrently , Especially when multiple transactions are trying to modify 、 When reading the same data , It's easy to have problems , Isolation is dealing with these problems .

Possible problems caused by concurrent transaction execution ：

1、 Dirty reading problem

      Business A Entering some data , It's business B Come and read the transaction A Data being entered , then B Go back and follow A Data input , If business A In the transaction B After reading the data, the data is modified , It's time for business B The data read may be a " Dirty data "（ This data is only a temporary result , Not the end result ）, This is the dirty reading problem .

How to deal with dirty reading ？

      Lock when writing data , etc. A Let's write it completely and revise it before B see , At this time, the isolation between transactions is improved , But at the same time, the concurrency is also reduced .

2、 Unrepeatable read problem

      Handle according to the above , wait until A Write it all B Look again , But if A stay B When reading, I modified the code , Now B Let's go back , The data is different again , This data cannot be read repeatedly , What should I do ？ At this time, simply lock the reading , stay A When writing data B Cannot read , stay B Reading the data A It can't be modified . At this time, the isolation of transactions is improved , Concurrency is reduced again .

3、 The problem of unreal reading

      The above two operations , Lock both reading and writing , Although its isolation is improved , But it reduces efficiency , As a result, it takes longer to read and write , At this time, in the process of writing or reading, transactions can change other tables or other rows of this table （ Although transactions have a series of locks when committing isolation , But the whole database is not locked ）, Now B When reading data, you will find that the amount of data has changed , There was only one A.java, Now there are more B.java, This is the unreal reading problem . In general, multiple queries are made during the execution of a transaction , The result set of each query is different （ More or less ）, This is also a special non repeatable problem .

How to solve it ？

Transaction serialization execution .（ Serialization is the execution of transactions one by one ）

From the above, we also know that isolation and concurrency cannot be achieved at the same time , To improve isolation , Concurrency must be reduced .

The above modifications have been made , At this time, the degree of isolation is the highest , The degree of concurrency is the lowest , The most reliable data , And the slowest .

According to the above summary ,MySQL Isolation level of transactions in , The following are provided :
1. read uncommitted： Allow reading uncommitted data , The highest degree of concurrency , Minimum isolation , Will introduce dirty reading 、 It can't be read repeatedly 、 The problem of unreal reading .
2. read committed： Only data after submission is allowed to be read , Equivalent to writing and locking , The degree of concurrency is reduced , The degree of isolation has increased a little , Solved dirty reading , Will introduce non repeatable reading and unreal reading .
3. repeatable read： It's equivalent to locking both read and write , The degree of concurrency is reduced again , Isolation has increased again , It solves dirty reading and non repeatability , Will introduce unreal reading .
4. serializable： Serialization , The lowest degree of concurrency ( Serial execution ), The highest degree of isolation , Solved dirty reading , Non repeatability , The problem of unreal reading , But the slowest execution .

🥬 Summary

Here we mainly know what the index is , Its applicable scenarios , What is the cost and the data structure behind the index , The four basic characteristics of transactions . That's what we have today , If you have any questions, you can leave a message in the comment area .