Golang encapsulates the packages involved in MySQL and the differences between sqlx and Gorm

One 、 Preface

       This is the notes before moving , Just used golang When , notice mysql The packaging part of , Always curious about why so many bags are used , for example ：

"database/sql"
"github.com/go-sql-driver/mysql"
"github.com/jmoiron/sqlx"
"github.com/jmoiron/sqlx/reflectx"
"vitess.io/vitess/go/vt/sqlparser"
"github.com/Masterminds/squirrel"
"github.com/lann/builder"

What's the difference between these bags ？ What exactly are we packaging mysql Key packages needed ？

Two 、 Parse various packages

Reference resources ：Go Language standard library learning database/sql—— Database management tools

1、database/sql

      sql Bag is go Officially provided operation mysql An extension of , It mainly provides operation mysql Methods , For example, add, delete, change and check , Functions such as transaction .

sql The main way of operation depends on native sql To write , for example ：

db.Query("SELECT * FROM student")

2、go-sql-driver/mysql"

       This bag is also go Official mysql Drive pack . Connect mysql The steps are as follows ：

1. Import database engine , That is to say, import mysql drive 
2. Connect to database ： root:[email protected]tcp(host:port)/database?charset=utf8&parseTime=True&loc=Local
3. Set related parameters 
4. Test connection 

       The driver package here , Mainly import mysql drive , Let's have sql Packages can connect to databases , The main usage is as follows ：

_ "github.com/go-sql-driver/mysql"
db, err := sql.Open("mysql",config)  // The first parameter mysql It's usage 

       Use “_” The operation refers to the package, and the export function in the package cannot be called through the package name , But just to simply call it init() function . After importing the database engine , You can go through sql Bag sql.open() Method to connect to the database .

3、sqlx package

Reference resources ：
sqlx file
sqlx Instructions

      sqlx yes Go A package , It's built in with excellent database/sql A set of extensions are provided on the package . in other words .sqlx It is also an operation mysql The standard library of .

since sql The bag has been officially issued , So why do people prefer to use sqlx Bao Er ？

（1） Security 
  database/sql  No validation of the query text will be attempted ; It is sent to the server as is with the encoding parameters .
   and sqlx Special parameter binding methods are provided sqlx.DB.Rebind(string) string, Support mysql question mark (?)
   Or named Prepared Statements, avoid SQL Injected security issues 
  
（2） facilitate 
   Relative to native sql Bag query,sqlx.Rowx The main extension of is StructScan, It can automatically scan the results 
   To the domain in the corresponding structure (fileld) in . hold SQL The result set of execution is transformed into a data structure (Struct、Maps、Slices).
  
（3）sqlx Of get and select Method 

  Get and Select It's a very time-saving extension . They put query And very flexible scan Combine grammar .
  Get and Select It will close after executing the query Rows, And any problem encountered during the execution phase will return an error .
   shortcoming ：Select The whole result will be put into memory at one time . If the query result is not limited to a specific size , So it's best to use Query/StructScan Iterative method .

sqlx Bag sql Mode of operation , Is also dependent on native sql To write ：

db.Query("SELECT * FROM student")

4、reflectx package

sqlx Packages have special reflection requirements . especially , It needs to ：

 Ability to map names to fields 
 Understand the embedded structure 
 Learn how to map names to fields with specific tags 
 User specified name  ->  Field mapping function 

reflectx The package mainly extends the original reflect package , To achieve the above purpose .

5、squirrel package and Builder package

       The main functions of these two packages are sql generator , Realize query construction , For chained calls . We know the original sql Packages need to be written manually sql Of , Programmers like to be lazy , It is best to input the initial letter and it can be completed automatically , Little by little chain calls all the way to the end .

（1）sqlx + squirrel Realization

squirrel yes go Of sql generator , Similar to the query constructor used before , Chain call can be implemented , Such as ：
sql, args, err := sq.
Insert("users").Columns("name", "age").
Values("moe", 13).Values("larry", sq.Expr("? + 5", 12)).
ToSql()

（2）builder package

Builder  Originally for  Squirrel Compiling , It is a smooth  SQL  generator . This may be  Builder  Best example of .
Builder The method chain helps you write fluent  DSL

6、vitess.io/vitess/go/vt/sqlparser package

Be careful ： This address is now 404 了

(1) vitess-sqlparser
  vitess-sqlparser  just Go Language SQL and DDL Parser ( from vitess and TiDB Support )
(2)vitess
    Reference resources ：https://blog.51cto.com/u_15080030/2642755
  Vitess Is a tool for deploying 、 Expand and manage large mysql The database solution of instance cluster .

       The above understanding of the common mysql Related to the package , No wonder the big guys sealed mysql You need to introduce so many packages when you need to , Interlocking , From security to chain call , The combination and packaging make it easier for everyone to use .

3、 ... and 、sqlx and gorm

1.gorm package

Reference resources ：Gorm Official documents

GORM yes GoLang The best in ORM frame , Support MySQL、PostgreSQL、Sqlite、SQL Server, The function is very 
 Powerful , It can also be executed directly SQL And get the result set . The main operation mode is orm In the form of , be relative to sql For expansion packs ,
orm Is a deeper level of packaging , For simplicity sql, Use orm It can effectively improve efficiency .

orm yes object Relational Mapping abbreviation ,object It's the object ,relational It's the relationship ,mapping Mapping .
 Object relational mapping looks quite abstract , you 're right , It is an abstract concept . From the perspective of programming syntax ,
 It is in the specific operation of business objects , There is no need to go and complex SQL Deal with sentences , Just a simple 
 Manipulate the properties and methods of the object .

gorm It is also similar to other languages ORM How to operate the database , More inclined to all objects .

2、gorm and sqlx Comparison of

（1）sqlx Is a library , It allows you to scan the entire line into your structural variables . although sqlx Reduced for building  CRUD  And the number of typical lines written , But eventually you will still write duplicate code many times .
 Use  ORM  Can help reduce it and focus on your business logic .
（2） For complex sq, For example, write sub queries of reports , Multi table association operation ,orm The implementation may be more complex , Use native sql It's better .
（3） Efficiency 
     Reference resources ：https://tech.wmzhe.com/article/97958.html , obviously sqlx More efficient 
（4）orm Is a class mapped to a table , Fields are mapped to member variables . When there are many fields in the table ,orm Classes will be very verbose .

       It doesn't matter sqlx still gorm, It is enough for our daily development , Bloggers have stayed in several companies with sqlx Mainly ,ORM Although convenient , But more abstract , Maybe you think sqlx It's easier to get started .

       Bloggers think , The important thing is not to use sqlx Or use it gorm, It is important to unify standards and sql Encapsulation of layers . Unified standards are for the readability and maintainability of code ,sql Layer encapsulation is to prevent code redundancy and facilitate the insertion of onion model middleware , Achieve relevant indicators of service stability .

3、 encapsulation sql What are the concerns

（1） Read the central configuration , Connect mysql, such as apollo etc. 
（2） Add onion model , Such as adding trace,spanID,requestId Evenly distributed links are traced to context,
 add to log Print key logs 

end