shorm

shorm is a orm tool wriitted in golang, which supports database sharding and master-slave mode

Installation

go get github.com/shengzhi/shorm

Features

• Struct -> table mapping support
• Fluent API
• Transaction support
• Both ORM and raw sql support
• Database cluster support
• Master-slave mode support
• Struct field support, will be stored as json string

Drivers support

• MsSql: github.com/denisenkom/go-mssqldb
• Mysql: github.com/go-sql-driver/mysql
• MyMysql: github.com/ziutek/mymysql/godrv
• Postgres: github.com/lib/pq
• SQLite: github.com/mattn/go-sqlite3

Change log

Quick start

• Create engine from code

    cluster := &Cluster{TotalGroups: 5}
	cluster.Groups = append(cluster.Groups,
		&DbGroup{
			IsDefault: true,
			RangeFrom: 0,
			RangeTo:   3,
			Name:      "Group1",
			Nodes: []*DbNode{
				{
					Name:    "G1_Master",
					ConnStr: connstr,
					Type:    NodeType_Master,
				},
				{
					Name:    "G1_Node1",
					ConnStr: connstr,
					Type:    NodeType_Slave,
				},
			},
		},
	)
	cluster.Groups = append(cluster.Groups,
		&DbGroup{
			RangeFrom: 3,
			RangeTo:   5,
			Name:      "Group2",
			Nodes: []*DbNode{
				{
					Name:    "G2_Master",
					ConnStr: connstr2,
					Type:    NodeType_Master,
				},
				{
					Name:    "G2_Node1",
					ConnStr: connstr2,
					Type:    NodeType_Slave,
				},
			},
		})
	engine := shorm.NewEngine("mssql", &cluster)

• if no sharding value, sql query will be executed against all groups

• if the sharding value falls in[0,3), the sql query will be exeucted against Group1

• if the sharding value falls in[3,5), the sql query will be exeucted against Group2

• Create engine from config file

	engine, err := shorm.NewEngineFromConfig("cluster_config.json")

• Define go struct to mapping data table

	Create table T_User(
		UserId bigint not null,
		UserName varchar(50) not null default '',
		Age int not null default 0,
		CreatedTime datetime not null default getdate(),
	)
	Alter table T_User create cluster index PK_User(UserId asc)

	type User struct{
		TabName shorm.TableName `shorm:"T_User"`
		Id int64 `shorm:"UserId,pk,shard"`
		UserName string
		Age int32
		CreatedTime time.Time
	}

	func (u User) GetShardValue() int64{
		return u.Id
	}

• Get value via specify sharding value

	s := engine.StartSession()
	defer engine.EndSession(s)
	u := User{}
	s.ShardValue(1).Id(1).Get(&u)

it will generate sql as below:

	select top 1 * from T_User where UserId=1

and will locate the database "G1_node1"

• shorm provides three ways to locate the database

  • Specify sharding value by call session.ShardValue(value int64), with highest priority
  • Go struct implements Shardinger interface, with higher priority
  • Mark field with "shard" in shorm struct tag, with lowest priority

• Query operations

  • Omit columns

     	s.Omit("CreatedTime").Get(&u)

  • Include specified columns, ignore others

   	s.Cols("UserId,UserName").Get(&u)

  • Query multiple records against all groups

   	var slice []*User
   	s.Where("Age > ?", 18).Find(&slice)

  • Paging, if sharding value not be specified, shorm will query data from all groups, so struct collection must implement the sort interface to achieve pagination

   	type AscendingUser []*User
  
   	func(list AscendingUser) Len() int{
   		return len(list)
   	}
  
   	func(list AscendingUser) Less(i, j int) bool{
   		return list[i].Id < list[j].Id
   	}
  
   	func (list AscendingUser) Swap(i, j int){
   		list[i],list[j] = list[j],list[i]
   	}
  
   	minimumId := 0
   	pageSize := 10
   	pageIndex := 0
   	var slice AscendingUser
   	s.Where("UserId>?",minimumId).And("Age>?",18).Limit(pageIndex * pageSize, pageSize).Find(&slice)
   	sort.Sort(slice)
   	return slice[:pageSize]
   	================================================================
   	type DescendingUser []*User
  
   	func(list DescendingUser) Len() int{
   		return len(list)
   	}
  
   	func(list DescendingUser) Less(i, j int) bool{
   		return list[i].Id > list[j].Id
   	}
  
   	func (list DescendingUser) Swap(i, j int){
   		list[i],list[j] = list[j],list[i]
   	}
  
   	maximumId := math.MaxInt64
   	pageSize := 10
   	pageIndex := 0
   	var slice AscendingUser
   	s.Where("UserId<?",maximumId).And("Age>?",18).OrderBy("UserId desc").Limit(pageIndex * pageSize, pageSize).Find(&slice)
   	sort.Sort(slice)
   	return slice[:pageSize]

  • In operatiion

   	s.ShardValue(100).In("UserId",12,234,345).Find(&slice)

  • Between operation

   	s.ShardValue(100).Between("UserId",10,100).Find(&slice)

• Insert data

  • Insert one record

   	user := User{
   		Id: 1001,
   		UserName: "Test",
   		Age: 18,
   		CreatedTime: time.Now(),
   	}
   	count,err := s.Insert(&user)

  • Insert slice

   	users := make(*User, 0)
   	result, err := s.InsertSlice(&users)
   	fmt.Println(result.Success)
   	fmt.Println(result.FailedData)

  • Database transaction(only support single database transaction)

   	trans, err := engine.BeginTrans(shardValue)
   	if _, err = trans.Insert(&user1);err != nil{
   		trans.Rollback()
   		return
   	}
   	if _, err = trans.InsertSlice(&orderList);err != nil{
   		trans.Rollback()
   		return
   	}
   	trans.Commit()