Database -- acid of transaction -- introduction / explanation

brief introduction

        This article introduces the transaction of database ACID.

         The transaction is ACID Four features ： Atomicity (Atomic)、 Uniformity (Consistent)、 Isolation, (Isolation)、 persistence (Duration).

Atomicity (Atomic)

explain

Transaction must be atomic unit of work , Modify the data , All or nothing , All or nothing .

Uniformity (Consistent)

explain

The transaction is completed , All data must be kept consistent .

Example

         user A And the user B The sum of the two is 700, No matter A and B How to transfer money between banks , The money of two users must add up to 700, This is consistency .

Isolation, (Isolation)

explain

Changes made by concurrent transactions must be isolated from changes made by any other concurrent transactions .

Detailed explanation

         In a concurrent environment , When different transactions manipulate the same data at the same time , Each transaction has its own full data space . Changes made by concurrent firms must be isolated from changes made by any other concurrent firms .

         When transactions view data updates , The state of the data is either the state of another transaction before it was modified , Or another transaction changes its state , Transactions do not see data in the intermediate state .

persistence (Duration)

explain

After the transaction completes , The status change of the corresponding data in the database should be Permanence Of , It will not be rolled back .

Detailed explanation

         If the transaction has been committed , But data is also persistent , At this time, the system is powered off or down , You can also re operate the non persistent data according to the transaction log , Put it recovery To the state that the transaction ends successfully .

Pre written logs （Write-ahead logging, abbreviation WAL） Mechanism

         Many databases use pre written logs （Write-ahead logging, abbreviation WAL） Mechanism , To ensure transaction persistence and data integrity , At the same time, the frequent data refresh based on transactions is avoided to a great extent IO Impact on performance .

         In the use of WAL In the system , All changes are written to the log first , Then it is applied to the system state . Let's say a program loses power while it's doing something . On reboot , The program may need to know whether the operation performed at that time was successful or partially successful or failed . Program can check log file , Compare the planned operation content with the actual operation content in case of sudden power failure . On the basis of this comparison , The program can then decide whether to undo the action or continue to complete the action , Or keep it as it is .​