[MySQL database] index and transaction (often used in interview)

Indexes

1. What is index

The index is equivalent to the catalogue of the book , The main function is Improve search efficiency .

2. How to improve search efficiency

When searching from the database , For example, according to the conditions id=3 lookup ： You can traverse the table to query , But it's less efficient . How to improve efficiency ？

This requires finding ways to avoid traversal , The characteristics of some records can be represented by some special data structures , Use these features to reduce the number of comparisons , Speed up the comparison .

When the search efficiency is improved , It will also pay some price ：

It's equivalent to adding some paper to the book , namely Adding indexes will consume a certain amount of storage space , More data , The more space is consumed

The catalogue of the book identifies , Every time the content of the book is adjusted later , May affect the accuracy of the directory , You need to readjust the directory , Empathy , The index of the database is the same , When adding, deleting, checking and modifying , It is often necessary to adjust the index structure synchronously

3. The advantages and disadvantages of indexing

Brought by index benefits ： Improve the search speed
The index brings Disadvantage ： Takes up more space , It slows down the speed of adding, deleting, checking and modifying

Be careful ：

Although the disadvantages of index seem to be more , But index is still a very common thing , Because in daily use , Search operations are often more frequent .

4. Some basic operations of index

1. Look at the index ：show index from Table name
   Some constraints are automatically indexed , such as ：primary ,unique
2. Create an index for the specified column
   create index Index name on Table name ( Name );
   for example ：create index class_index on student(class);-- to class This column is indexed
   Be careful ： Creating an index is a very inefficient thing , Especially when there are a lot of data in the current table
   therefore , When you target online databases , If the table has no index , Don't rush to add an index …
3. Delete index
   drop index Index name on student Table name
   for example ：drop index class_index on student;
   Be careful ： Deleting an index is the same as creating an index , They are all inefficient operations , It's also easy to hang up the database
   therefore , When creating tables , You should plan the index .

5. Data structure behind index ( a key )

To make the search faster , The most important thing is to reduce access disk IO The number of times , Because of entering disk IO More time than Compare It takes much longer .

therefore , use B+ Trees Structure , Here's why ：

1. Use B+ When searching the tree , Access to the IO The number of times is relatively small
2. be-all Queries will eventually fall on leaf nodes , Every time I look up IO The times are almost the same , The query speed is stable
3. After the leaf node is linked with a linked list , very Suitable for range finding
4. All data storage （ load ） They are all placed on leaf nodes , Non leaf nodes are only saved key that will do , Therefore, non leaf nodes occupy less space as a whole , It can even be cached in memory , This makes Access disk IO The number of times is almost gone , Greatly speed up the query speed .

The reason why binary search trees and red black trees are not used is ： The height of the tree is too high , Lead to IO More visits .
Reasons for not using hash tables ： Cannot support range lookup .

Business

1. What is business

Transaction is a function born to package several independent operations into a whole .
Transactions are atomic ： Execution time , Or it can be implemented as a whole , Or none at all , That is, inseparable .

2. Application scenarios of transactions

such as ：A——>B Transfer accounts 500
A Balance of -500

update  Account form  set balance = balance-500 where name = "A";

B Balance of +500

update  Account form  set balance = balance+500 where name = "B";

When performing this transfer operation , There may be a problem ： When A Turn around 500 when ,B I haven't received , But the machine broke down .
Obviously, this state is unscientific , But the use of The atomicity of transactions Can solve this problem

3. How to guarantee transaction atomicity

In execution to the second SQL Before , It is impossible to predict that this implementation will fail （ Suppose that the second execution will fail ）
therefore , When execution fails , There are databases that perform some “ Restore ” operation , To eliminate the front SQL The impact . This reductive operation , be called “ Roll back ”.
So it looks ： When execution fails , None of them .

4. Use of transactions

1. Open transaction ：start transaction;
2. Execute more than one SQL
3. Roll back or commit ：rollback ,commit

explain ：rollback It's all failure ,commit It's all success

5. The four characteristics of business ：

1） Atomicity ：
2） Uniformity ： Before and after transaction execution , The data in the database must be legal . for example ： After the transfer , You cannot have a negative account 
3） persistence ： Once the transaction is committed , It is persistent and stored , The data is stored in the hard disk .
4） Isolation, ： It's important , Let's introduce it separately 

6. Graphic dirty reading 、 It can't be read repeatedly 、 Fantasy reading

Deal with dirty reading ： In the process of writing , Others can't read （ Locked state ） After the modification , Others can read （ Unlock ）

Processing non repeatable ： Lock the reader , I can't write when I read , It solves the problem of non repeatable reading .

Deal with unreal reading ： Serialization completes the transaction , namely ： Write ——> read ——> Write ——> read

7. Isolation of transactions

Isolation describes , When transactions are concurrent , What happened . When multiple transactions are executed concurrently , In particular, many transactions try to modify / When reading the same data , It's easy to have some problems , Transaction isolation can solve this problem .

Between transactions concurrency and Isolation, You can't have both ：

When concurrency is high , Poor isolation ; When concurrency is low , High isolation . The balance between the two depends on the needs of the characteristics of the transaction .
The higher the concurrency , The faster transactions are executed ; The higher the isolation , The result of executing transactions is more accurate .

8. MySQL Isolation level of transactions in

1. read uncommitted： Allow reading of data submitted for , The highest degree of concurrency , Minimum isolation , Will introduce dirty reading + It can't be read repeatedly + The problem of unreal reading .
2. read ncommitted： Only the submitted data is allowed to be read , It is equivalent to locking the write , The degree of concurrency is reduced , The degree of isolation has increased a little , Solved dirty reading , Will introduce non repeatable + The problem of 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 , Solve the problem of dirty reading and non repeatable reading , Will introduce unreal reading
4. serialization： Serialization （ After one execution , To execute the next ）, The highest degree of isolation , Solved dirty reading + It can't be read repeatedly + The problem of unreal reading .

When the isolation level needs to be adjusted ： It can be modified by my.ini This configuration file , To set the current isolation level , According to the actual needs , Adjust the isolation level .