package main
import (
_ "github.com/lib/pq"
"github.com/xormplus/xorm"
"fmt"
"time"
"strconv"
)
//用户表结构
type User struct {
Id int `xorm:"not null pk autoincr INTEGER"`
Name string `xorm:"VARCHAR(20)"`
Created time.Time `xorm:"default 'now()' DATETIME"`
ClassId int `xorm:"default 1 INTEGER"`
}
//Class表结构
type Class struct {
Id int `xorm:"not null pk autoincr INTEGER"`
Name string `xorm:"VARCHAR(20)"`
}
//临时表结构
type UserClass struct{
User `xorm:"extends"`
Name string
}
//此方法仅用于orm查询时,查询表认定
func (UserClass) TableName() string {
return "public.user"
}
func main() {
//1.创建db引擎
db, err := xorm.NewPostgreSQL("postgres://postgres:123@localhost:5432/test?sslmode=disable")
if err != nil {
fmt.Println(err)
}
//2.显示sql语句
db.ShowSQL(true)
//3.设置连接数
db.SetMaxIdleConns(2000)
db.SetMaxOpenConns(1000)
cacher := xorm.NewLRUCacher(xorm.NewMemoryStore(), 500)//缓存的条数
db.SetDefaultCacher(cacher)
//4.执行插入语句的几种方式
//4.1 orm插入方式:不好控制,如果仅仅插入的对象的属性是name='ftq',那么其他的零值会一同insert,orm方式对零值的处理有点不太好
if false {
user := new(User)
user.Name = "ftq"
_, err = db.Insert(user)
if err != nil {
fmt.Println(err)
}
}
//4.2 命令插入方式
//4.2.1 db.Exec():单事务单次提交
if false{
sql:="insert into public.user(name) values(?)"
db.Exec(sql,"ft4")
}
//4.2.2 db.SQL().Execute():单事务准备了Statement处理sql语句
if false{
sql:="insert into public.user(name) values(?)"
db.SQL(sql,"ft5").Execute()
}
//4.2.3使用sql配置文件管理语句,两种载入配置的方式LoadSqlMap()和RegisterSqlMap(),以及SqlMapClient()替代SQL()
if false {
err = db.LoadSqlMap("./sql.xml")
//err = db.RegisterSqlMap(xorm.Xml("./","sql.xml"))
if err != nil {
fmt.Println(err)
}
db.SqlMapClient("insert_1","ft7").Execute()
}
//5.执行查询的几种方式
//5.1 orm查询:在user初始化的时候,该orm查询直接通过扫描user类型确定表名;组合使用Where(),Get()
if false {
user := new(User)
boolget ,err2 :=db.Where("id=?",5).Get(user)
fmt.Println(boolget,err2,user)
}
//5.2 orm查询:组合使用Where(),Get(),And()
if false {
user := new(User)
boolget ,err2 :=db.Where("id=?",5).And("name=?","ft7").And("id>?",3).Get(user)
fmt.Println(boolget,err2,user)
}
//5.3 orm查询: AllCols()查询所有列,Cols()查询部分列,Find()解析多行结果,Get()解析单行结果
if false {
users := new ([]User)
err = db.AllCols().Find(users)
//err = db.Cols("id","name").Find(users)
if err !=nil {
fmt.Println(err)
}
fmt.Println(users)
}
//5.4 orm查询:连接查询Join()
if false {
users := new([]UserClass)
db.Join("INNER","class","user.class_id=class.id").Find(users)
//db.SQL("select u.id,u.name,c.name from public.user as u left join public.class as c on u.class_id=c.id").Find(users)
fmt.Println(users)
}
//5.5 sql查询略
if false {
//和insert类似,Find查找多行结果,Get获取 单行结果
users := new([]UserClass)
db.SQL("select u.id,u.name,c.name from public.user as u left join public.class as c on u.class_id=c.id").Find(users)
fmt.Println(users)
}
//5.6 链式查找
if false {
//值得一提的是,支持查找某行的某个字段,不过一般在sql语句中就可以完成过滤,如果sql语句过于复杂,可以链式查找过滤
id := db.SQL("select * from public.user").Query().Results[0]["id"]
fmt.Println(id)
}
//6.执行更新
//6.1 ORM方式: 只有非0值的属性会被更新,user的id和created都是默认零值,不被处理
if false {
user :=new(User)
user.Name="ftx"
//[xorm] [info] 2018/02/08 12:04:01.330624 [SQL] UPDATE "user" SET "name" = $1 WHERE "id"=$2 []interface {}{"ftx", 4}
db.Id(4).Update(user)
}
//6.2 SQL方式略,和insert类似
//7.事务
//7.1简单事务
if false {
session :=db.NewSession()
defer session.Close()
session.Begin()
//业务:新添加学生,并且创建新的班级,如果班级因为主键冲突创建失败,则整个事务回滚
_,err =session.SQL("insert into public.user(name,class_id) values('ft13',2)").Execute()
//表中已经有id=3的班级了
_,err =session.SQL("insert into public.class(id,name) values(3,'高中3班')").Execute()
if err!=nil {
session.Rollback()
}
session.Commit()
}
//7.2嵌套事务
if false {
session := db.NewSession()
defer session.Close()
session.Begin()
_,err=session.Exec("insert into public.user(name,class_id) values('ft23',2)")
if err!=nil {
session.Rollback()
}
_,err=session.Exec("insert into public.user(id,name,class_id) values(1,'ft24',2)")
if err!=nil {
session.Rollback()
}
tx,_:=session.BeginTrans()
_,err=tx.Session().Exec("insert into public.user(name,class_id) values('ft25',2)")
if err!=nil {
tx.RollbackTrans()
}
tx.CommitTrans()
session.Commit()
}
//8.缓存:使用Raw方式修改以后,需要清理缓存
if true {
//建立500条数据
session := db.NewSession()
defer session.Close()
if false {
session.Begin()
for i := 30; i 530; i++ {
value := "ft" + strconv.Itoa(i)
_, err = session.Exec("insert into public.user(name) values(?)", value)
if err != nil {
session.Rollback()
}
}
session.Commit()
}
//查询前531条数据,并随意输出其中一条
users := make([]User,10)
db.SQL("select * from public.user where id531 order by id").Find(users)
fmt.Println("读第一遍:","id:",users[50].Id,"name:",users[50].Name)
db.SQL("select * from public.user where id531 order by id").Find(users)
fmt.Println("读第二遍:","id:",users[50].Id,"name:",users[50].Name)
var step int =1
stepString := users[50].Name + strconv.Itoa(step)
session.Exec("update public.user set name=? where id =45",stepString)
//清理缓存
db.ClearCache(new(User))
time.Sleep(5*time.Second)
session.SQL("select * from public.user where id531 order by id").Find(users)
fmt.Println("读第三遍:","id:",users[50].Id,"name:",users[50].Name)
//虽然很不好意思,但是就算开启了缓存数据也是脏了
}
//9.读写分离
if false {
//假设有多台服务器用来响应客户的读请求
var dbGroup *xorm.EngineGroup
conns :=[]string {
"postgres://postgres:123@localhost:5432/test?sslmode=disable",
"postgres://postgres:123@localhost:5432/test?sslmode=disable",
"postgres://postgres:123@localhost:5432/test?sslmode=disable",
"postgres://postgres:123@localhost:5432/test?sslmode=disable",
}
//负载均衡策略:(特性自行百度)
// 1.xorm.RandomPolicy()随机访问负载均衡,
// 2.xorm.WeightRandomPolicy([]int{2, 3,4})权重随机负载均衡
// 3.xorm.RoundRobinPolicy() 轮询访问负载均衡
// 4.xorm.WeightRoundRobinPolicy([]int{2, 3,4}) 权重轮训负载均衡
// 5.xorm.LeastConnPolicy()最小连接数负载均衡
dbGroup, err = xorm.NewEngineGroup("postgres", conns, xorm.RoundRobinPolicy())
//dbGroup使用方法和db一致
//简单查询
dbGroup.SQL("inser into public.users(name) values('ft2000')").Execute()
dbGroup.Exec("inser into public.users(name) values('ft2001')")
//事务查询
session :=dbGroup.NewSession()
defer session.Close()
session.Begin()
_,err = session.Exec("inser into public.users(name) values('ft2001')")
if err!=nil {
session.Rollback()
}
session.Commit()
}
}
//注意:
//1.postgresql好像不会默认按id增长排序,所以书写sql语句要提前写好order by id ,楼主没怎么写,咳咳
//2. [5.4] postgresql建表会建在public策略的table里,所以查询语句表明写的是public.xxxx,这也造成了连表orm查询会发生前缀报错,比如变成了"SELECT * FROM "public"."user" INNER JOIN "class" ON user.class_id=class.id 这和内部的split有关,
//3.[8.]带的缓存好像很容易失效,在创建500个数据后,经过查查改查的操作,查询到的结果是一样的始终是一样的,本来改值之后应该最后一遍查会变化,然而并没有,缓存功能即使清理了缓存,还是会读到脏的