[MySQL] MySQL Performance Optimization Practice: introduction of database lock and index search principle

Row-level locks & Table lock

What is a row level lock ？ What is a table level lock ？ Students who have learned the program must have a certain understanding of the concept of lock , As the name suggests, one is to lock a line , One is to lock a whole watch , Let's look at the following example ：

Typical cases

Existing table T_A The structure and data are as follows

Start a transaction （session_1） And execute the following statement

Business one

stay session_1 Without submitting , Start a new business （session_2） And execute the statement

Business two

As a result, the transaction timed out , The problem is coming. , An error is a lock timeout, indicating that a table level lock is used , But it doesn't mean InnoDB Is it a row lock ？ It should be noted that row level locks cannot take effect until indexes are used

We give fields column_2 Add index after , After performing the above operation again, it is found that the two transactions do not interfere with each other, and the execution is successful , In high concurrency scenarios, it can greatly improve the efficiency of the program .

Deadlock

Deadlock is mainly caused by different transactions competing for the same resource .

Specific cases

It is also the above T_A Update rows in the table , Deadlock caused when competing for resources

It can be seen that the deadlock will immediately report an error after it occurs . There is a mistake here , In our development, we often encounter too high concurrency, which leads to a large number of locks in the database that cannot be released in time, which is often described as deadlock by developers , It's not , The real deadlock will report an error at the moment of occurrence .

The query principle of index

In the last article, I made a brief introduction to the use of index , Some students may not be very clear about the difference between the two concepts of primary key and index , Let me briefly introduce how to find records through primary keys and indexes in the database ：

Primary key storage structure

Index storage structure

The above figure shows the storage structure of primary key and index respectively , You can see that the corresponding record can be found directly through the primary key search , To search by index, you need to find the primary key of the corresponding record , Then find the corresponding record according to the primary key .

Two conclusions can be drawn from the above analysis ：

1、 You have to set the primary key , Because the index ultimately needs to locate specific records through the primary key

2、 The primary key should preferably be numeric , In this way, the storage distribution of primary keys is more uniform and the search efficiency is higher