One 、 Four characteristics of transactions

The isolation level of the transaction is ACID, That is atomicity 、 Uniformity 、 Isolation, 、 persistence .

1、 Atomicity

Atomicity means that the operation of the whole transaction is either committed successfully , Either all failures roll back .

2、 Uniformity

Consistency means that the database always changes from one consistent state to another consistent state , If the transaction is not committed , The changes made will not be saved to the database .

3、 Isolation,

Changes made by a transaction are made before the final submission , Is invisible to other transactions .

4、 persistence

Once the transaction is committed , The changes made will be permanently saved to the database .

Realized ACID The database of , Usually stronger cpu Processing power and more memory and more disk space .

Two 、 If there is no isolation , What will happen ？

1、 Dirty reading

Business A	Business B
Open transaction A
	Open transaction B
The query balance is 100
	Increase balance to 150
The query balance is 150

Why is it dirty reading , Because no submission , If you roll back , Then the data read at this time is just a mirage .

2、 It can't be read repeatedly

Business A	Business B
Open transaction A
	Open transaction B
The query balance is 100
	Increase balance to 150
The query balance is 100
	Commit transaction
The query balance is 150

Refers to a transaction scope , Query a data multiple times , But got different results .
Between two reads in the first transaction , Due to the modification of the second transaction , The data read by the first transaction twice may be different .

3、 Fantasy reading

Business A	Business B
Open transaction A
	Open transaction B
Inquire about id<3 The records of 3 strip
	Add a id=2 The record of
	Commit transaction
Inquire about id<3 There are four records

Another example is business T1 A data item is removed from all rows in a table “1” It is amended as follows “2” The operation of , When the transaction T2 Another row of data item is inserted into this table, which is “1” The data of , And submit it to the database . And operating transactions T1 If the user views the data that has just been modified , Will you find the data or 1, Actually this row is from the transaction T2 Added in , It's like hallucinating , This is where the phantom reading occurs .
Unreal reading and non repeatable reading are both read and submitted data , But unreal reading is aimed at a batch of data , Non repeatable reading is for a row of data .

3、 ... and 、 The isolation level of the transaction

1、 Uncommitted read

Changes in transactions , Even if it's not submitted , It's also visible to other things , Transactions can read uncommitted data , Uncommitted reading cannot avoid any of the above problems .

2、 Read submitted

When a transaction begins , You can only see the changes made by the submitted firm . Avoid dirty reading .

3、 Repeatable

In the same business , The result of multiple reads is the same . It can avoid dirty reading and repeatable reading , But unreal reading cannot be avoided .

4、 Serialization

Is the highest isolation level , Force transactions to execute serially , Avoid unreal reading 、 It can't be read repeatedly 、 Dirty reading problem , The serial will lock each row of data , Therefore, it may cause a lot of timeout and lock contention problems . Only if the consistency of data is ensured and a large number of concurrency can be accepted , To consider using this level .