目錄
- 一、設置環境
- 1.1、安裝 SQL Server
- 1.2、安裝 GoLang
- 1.3、安裝 ODBC 驅動程序和 SQL 命令行實用工具 SQL 服務器
- 二、使用 SQL 服務器創建 Go 應用程序
- 2.1、創建連接到 SQL Server 并執行查詢的 Go 應用
- 2.2、創建一個使用 GORM 連接到 SQL Server 的 Go 應用程序
- 三、讓 Go 應用的速度提高 100 倍
- 3.1、使用 sqlcmd 創建一個包含 5 萬個的新表
- 3.2、創建一個 Go 應用程序，用于查詢此表并測量所花費的時間
- 3.3、測量運行查詢所需的時間
- 3.4、使用 SQLCMD 向表中添加列存儲索引
- 3.5、重新運行 columnstore.go 腳本，并注意這次完成查詢所花費的時間
- 總結
一、設置環境
在 Ubuntu 機器上安裝 SQL Server 2017和安裝運行 GoLang 所需的依賴項。
1.1、安裝 SQL Server
為了確保 SQL Server 的最佳性能，計算機應至少具有 4 GB 的內存。
（2）安裝 SQLCMD 后，可以使用以下命令連接到 SQL Server：
sqlcmd -S localhost -U sa -P yourpassword 1> # You're connected! Type your T-SQL statements here. Use the keyword 'GO' to execute each batch of statements.
（3）測試數據庫。結果將打印到標準輸出。
sqlcmd -S localhost -U sa -P yourpassword -Q "SELECT @@VERSION"
--------------------------------------------------------
Microsoft SQL Server vNext (CTP2.0) - 14.0.500.272 (X64)
?? ?Apr 2 2023 11:44:40
?? ?Copyright (c) Microsoft Corporation
? ? on Linux (Ubuntu 16.04)1 rows(s) returned
Executed in 1 ns
至此，已成功在 Ubuntu 機器上安裝 SQL Server 命令行實用程序，已經在 Ubuntu 計算機上成功安裝并設置 GoLang 和 mssql-tools。現在擁有開始使用 SQL Server 編寫 Go 應用程序所需的一切。
二、使用 SQL 服務器創建 Go 應用程序
安裝 SQL Server 和 GoLang 后，現在可以繼續創建新的 Go 項目。在這里，將探討三個簡單的應用程序。其中一個將連接并打印數據庫服務器的SQL Server版本，另一個將執行基本的插入，更新，刪除和選擇操作，第三個將使用GORM，一種流行的對象關系映射（ORM）框架，用于Go執行相同的操作。
2.1、創建連接到 SQL Server 并執行查詢的 Go 應用
（1）創建新的項目目錄并安裝 Go 依賴項。
cd ~/ #Create Project Directory mkdir SqlServerSample cd SqlServerSample # Get and install the SQL Server driver for Go go get github.com/denisenkom/go-mssqldb go install github.com/denisenkom/go-mssqldb
（2）通過使用 sqlcmd 連接到 SQL Server 并執行以下命令，創建將用于本教程其余部分的數據庫。不要忘記使用自己的用戶名和密碼更新用戶名和密碼。
sqlcmd -S 127.0.0.1 -U sa -P <你的> -Q "CREATE DATABASE SampleDB;"
（3）創建一個連接到 SQL Server 的簡單 Go 應用。
在 SqlServerSample 文件夾中創建一個名為 connect.go 的文件。將以下內容復制并粘貼到文件中。不要忘記使用自己的用戶名和密碼更新用戶名和密碼。
此示例使用 GoLang 上下文方法來確保存在與數據庫服務器的活動連接。
package main
import (
_ "github.com/denisenkom/go-mssqldb"
"database/sql"
"context"
"log"
"fmt"
)
// Replace with your own connection parameters
var server = "localhost"
var port = 1433
var user = "sa"
var password = "xxxxxx"
var db *sql.DB
func main() {
var err error
// Create connection string
connString := fmt.Sprintf("server=%s;user id=%s;password=%s;port=%d",
server, user, password, port)
// Create connection pool
db, err = sql.Open("sqlserver", connString)
if err != nil {
log.Fatal("Error creating connection pool: " + err.Error())
}
log.Printf("Connected!\n")
// Close the database connection pool after program executes
defer db.Close()
SelectVersion()
}
// Gets and prints SQL Server version
func SelectVersion(){
// Use background context
ctx := context.Background()
// Ping database to see if it's still alive.
// Important for handling network issues and long queries.
err := db.PingContext(ctx)
if err != nil {
log.Fatal("Error pinging database: " + err.Error())
}
var result string
// Run query and scan for result
err = db.QueryRowContext(ctx, "SELECT @@version").Scan(&result)
if err != nil {
log.Fatal("Scan failed:", err.Error())
}
fmt.Printf("%s\n", result)
}
（4）運行應用程序。
go run connect.go
執行結果：
Connected!
Microsoft SQL Server 2017 (CTP2.1) - 14.0.600.250 (X64)
? ? ? ? Apr 2 2017 12:21:23
? ? ? ? Copyright (C) 2017 Microsoft Corporation. All rights reserved.
? ? ? ? Developer Edition (64-bit) on Linux (Ubuntu 16.04.2 LTS)
（5）在 SqlServerSample 文件夾中創建一個名為 CreateTestData 的文件.sql。將以下 T-SQL 代碼復制并粘貼到其中。這將創建一個架構、表并插入幾行。
CREATE SCHEMA TestSchema; GO CREATE TABLE TestSchema.Employees ( Id INT IDENTITY(1,1) NOT NULL PRIMARY KEY, Name NVARCHAR(50), Location NVARCHAR(50) ); GO INSERT INTO TestSchema.Employees (Name, Location) VALUES (N'Jared', N'Australia'), (N'Nikita', N'India'), (N'Tom', N'Germany'); GO SELECT * FROM TestSchema.Employees; GO
（6）使用 sqlcmd 連接到數據庫并運行 SQL 腳本以創建架構、表并插入一些行。
sqlcmd -S 127.0.0.1 -U sa -P <你的> -d SampleDB -i ./CreateTestData.sql
執行結果：
CREATE SCHEMA TestSchema; Executed in 0 ms CREATE TABLE TestSchema.Employees ( Id INT IDENTITY(1,1) NOT NULL PRIMARY KEY, Name NVARCHAR(50), Location NVARCHAR(50) ); Executed in 0 ms INSERT INTO TestSchema.Employees (Name, Location) VALUES (N'Jared', N'Australia'), (N'Nikita', N'India'), (N'Tom', N'Germany'); Executed in 0 ms SELECT * FROM TestSchema.Employees; Id Name Location -- ------ --------- 1 Jared Australia 2 Nikita India 3 Tom Germany 3 row(s) returned Executed in 1 ms
（7）在 SqlServerSample 文件夾中創建一個名為 crud.go 的新文件。將以下代碼復制并粘貼到其中。這將插入、更新、刪除和讀取幾行。
package main
import (
_ "github.com/denisenkom/go-mssqldb"
"database/sql"
"context"
"log"
"fmt"
"errors"
)
var db *sql.DB
var server = "localhost"
var port = 1433
var user = "sa"
var password = "你的"
var database = "SampleDB"
func main() {
// Build connection string
connString := fmt.Sprintf("server=%s;user id=%s;password=%s;port=%d;database=%s;",
server, user, password, port, database)
var err error
// Create connection pool
db, err = sql.Open("sqlserver", connString)
if err != nil {
log.Fatal("Error creating connection pool: ", err.Error())
}
ctx := context.Background()
err = db.PingContext(ctx)
if err != nil {
log.Fatal(err.Error())
}
fmt.Printf("Connected!\n")
// Create employee
createID, err := CreateEmployee("Jake", "United States")
if err != nil {
log.Fatal("Error creating Employee: ", err.Error())
}
fmt.Printf("Inserted ID: %d successfully.\n", createID)
// Read employees
count, err := ReadEmployees()
if err != nil {
log.Fatal("Error reading Employees: ", err.Error())
}
fmt.Printf("Read %d row(s) successfully.\n", count)
// Update from database
updatedRows, err := UpdateEmployee("Jake", "Poland")
if err != nil {
log.Fatal("Error updating Employee: ", err.Error())
}
fmt.Printf("Updated %d row(s) successfully.\n", updatedRows)
// Delete from database
deletedRows, err := DeleteEmployee("Jake")
if err != nil {
log.Fatal("Error deleting Employee: ", err.Error())
}
fmt.Printf("Deleted %d row(s) successfully.\n", deletedRows)
}
// CreateEmployee inserts an employee record
func CreateEmployee(name string, location string) (int64, error) {
ctx := context.Background()
var err error
if db == nil {
err = errors.New("CreateEmployee: db is null")
return -1, err
}
// Check if database is alive.
err = db.PingContext(ctx)
if err != nil {
return -1, err
}
tsql := "INSERT INTO TestSchema.Employees (Name, Location) VALUES (@Name, @Location); select convert(bigint, SCOPE_IDENTITY());"
stmt, err := db.Prepare(tsql)
if err != nil {
return -1, err
}
defer stmt.Close()
row := stmt.QueryRowContext(
ctx,
sql.Named("Name", name),
sql.Named("Location", location))
var newID int64
err = row.Scan(&newID)
if err != nil {
return -1, err
}
return newID, nil
}
// ReadEmployees reads all employee records
func ReadEmployees() (int, error) {
ctx := context.Background()
// Check if database is alive.
err := db.PingContext(ctx)
if err != nil {
return -1, err
}
tsql := fmt.Sprintf("SELECT Id, Name, Location FROM TestSchema.Employees;")
// Execute query
rows, err := db.QueryContext(ctx, tsql)
if err != nil {
return -1, err
}
defer rows.Close()
var count int
// Iterate through the result set.
for rows.Next() {
var name, location string
var id int
// Get values from row.
err := rows.Scan(&id, &name, &location)
if err != nil {
return -1, err
}
fmt.Printf("ID: %d, Name: %s, Location: %s\n", id, name, location)
count++
}
return count, nil
}
// UpdateEmployee updates an employee's information
func UpdateEmployee(name string, location string) (int64, error) {
ctx := context.Background()
// Check if database is alive.
err := db.PingContext(ctx)
if err != nil {
return -1, err
}
tsql := fmt.Sprintf("UPDATE TestSchema.Employees SET Location = @Location WHERE Name = @Name")
// Execute non-query with named parameters
result, err := db.ExecContext(
ctx,
tsql,
sql.Named("Location", location),
sql.Named("Name", name))
if err != nil {
return -1, err
}
return result.RowsAffected()
}
// DeleteEmployee deletes an employee from the database
func DeleteEmployee(name string) (int64, error) {
ctx := context.Background()
// Check if database is alive.
err := db.PingContext(ctx)
if err != nil {
return -1, err
}
tsql := fmt.Sprintf("DELETE FROM TestSchema.Employees WHERE Name = @Name;")
// Execute non-query with named parameters
result, err := db.ExecContext(ctx, tsql, sql.Named("Name", name))
if err != nil {
return -1, err
}
return result.RowsAffected()
}
（8）運行 crud.go 應用以查看結果。
go run crud.go
執行結果：
Connected!
Inserted ID: 4 successfully.
ID: 1, Name: Jared, Location: Australia
ID: 2, Name: Nikita, Location: India
ID: 3, Name: Tom, Location: Germany
ID: 4, Name: Jake, Location: United States
Read 4 row(s) successfully.
Updated 1 row(s) successfully.
Deleted 1 row(s) successfully.
2.2、創建一個使用 GORM 連接到 SQL Server 的 Go 應用程序
（1）創建應用目錄并初始化 Go 依賴項。
cd ~/ mkdir SqlServerGormSample cd SqlServerGormSample # Get and install the SQL Server driver for Go go get github.com/denisenkom/go-mssqldb go install github.com/denisenkom/go-mssqldb
（2）將以下內容粘貼到名為orm.go的文件中。確保將密碼變量替換為您自己的變量。
package main
import (
"fmt"
"github.com/jinzhu/gorm"
_ "github.com/jinzhu/gorm/dialects/mssql"
"log"
)
var server = "localhost"
var port = 1433
var user = "sa"
var password = "你的"
var database = "SampleDB"
// Define a User model struct
type User struct {
gorm.Model
FirstName string
LastName string
}
// Define a Task model struct
type Task struct {
gorm.Model
Title string
DueDate string
IsComplete bool
UserID uint
}
// Read and print all the tasks
func ReadAllTasks(db *gorm.DB){
var users []User
var tasks []Task
db.Find(&users)
for _, user := range users{
db.Model(&user).Related(&tasks)
fmt.Printf("%s %s's tasks:\n", user.FirstName, user.LastName)
for _, task := range tasks {
fmt.Printf("Title: %s\nDueDate: %s\nIsComplete:%t\n\n",
task.Title, task.DueDate, task.IsComplete)
}
}
}
// Update a task based on a user
func UpdateSomeonesTask(db *gorm.DB, userId int){
var task Task
db.Where("user_id = ?", userId).First(&task).Update("Title", "Buy donuts for Luis")
fmt.Printf("Title: %s\nDueDate: %s\nIsComplete:%t\n\n",
task.Title, task.DueDate, task.IsComplete)
}
// Delete all the tasks for a user
func DeleteSomeonesTasks(db *gorm.DB, userId int){
db.Where("user_id = ?", userId).Delete(&Task{})
fmt.Printf("Deleted all tasks for user %d", userId)
}
func main() {
connectionString := fmt.Sprintf("server=%s;user id=%s;password=%s;port=%d;database=%s",
server, user, password, port, database)
db, err := gorm.Open("mssql", connectionString)
if err != nil {
log.Fatal("Failed to create connection pool. Error: " + err.Error())
}
gorm.DefaultCallback.Create().Remove("mssql:set_identity_insert")
defer db.Close()
fmt.Println("Migrating models...")
db.AutoMigrate(&User{})
db.AutoMigrate(&Task{})
// Create awesome Users
fmt.Println("Creating awesome users...")
db.Create(&User{FirstName: "Andrea", LastName: "Lam"}) //UserID: 1
db.Create(&User{FirstName: "Meet", LastName: "Bhagdev"}) //UserID: 2
db.Create(&User{FirstName: "Luis", LastName: "Bosquez"}) //UserID: 3
// Create appropriate Tasks for each user
fmt.Println("Creating new appropriate tasks...")
db.Create(&Task{
Title: "Do laundry", DueDate: "2017-03-30", IsComplete: false, UserID: 1})
db.Create(&Task{
Title: "Mow the lawn", DueDate: "2017-03-30", IsComplete: false, UserID: 2})
db.Create(&Task{
Title: "Do more laundry", DueDate: "2017-03-30", IsComplete: false, UserID: 3})
db.Create(&Task{
Title: "Watch TV", DueDate: "2017-03-30", IsComplete: false, UserID: 3})
// Read
fmt.Println("\nReading all the tasks...")
ReadAllTasks(db)
// Update - update Task title to something more appropriate
fmt.Println("Updating Andrea's task...")
UpdateSomeonesTask(db, 1)
// Delete - delete Luis's task
DeleteSomeonesTasks(db, 3)
}
（3）運行 orm.go 應用。
go run orm.go
執行結果：
[info] removing callback `mssql:set_identity_insert` from C:/Projects/golang-experiments/tutorials/orm.go:70
Migrating models...
Creating awesome users...
Creating new appropriate tasks...Reading all the tasks...
Andrea Lam's tasks:
Title: Do laundry
DueDate: 2017-03-30
IsComplete:falseMeet Bhagdev's tasks:
Title: Mow the lawn
DueDate: 2017-03-30
IsComplete:falseLuis Bosquez's tasks:
Title: Do more laundry
DueDate: 2017-03-30
IsComplete:falseTitle: Watch TV
DueDate: 2017-03-30
IsComplete:falseUpdating Andrea's task...
Title: Buy donuts for Luis
DueDate: 2017-03-30
IsComplete:falseDeleted all tasks for user 3
三、讓 Go 應用的速度提高 100 倍
已了解基礎知識，接下來可以了解如何使用 SQL Server 改進應用。通過列存儲索引的簡單示例，以及它們如何提高數據處理速度。與傳統行存儲索引相比，列存儲索引在分析工作負荷上可實現高達 100 倍的性能，并將數據壓縮提高多達 10 倍。
3.1、使用 sqlcmd 創建一個包含 5 萬個的新表
（1）切換到主目錄并為項目創建一個文件夾。
cd ~/mkdir SqlServerColumnstoreSamplecd SqlServerColumnstoreSample
（2）在 SqlServerColumnstoreSample 文件夾中創建一個名為 CreateSampleTable 的新文件.sql文件。將下面的 T-SQL 代碼粘貼到新的 SQL 文件中。保存并關閉文件。
WITH a AS (SELECT * FROM (VALUES(1),(2),(3),(4),(5),(6),(7),(8),(9),(10)) AS a(a)) SELECT TOP(5000000) ROW_NUMBER() OVER (ORDER BY a.a) AS OrderItemId ,a.a + b.a + c.a + d.a + e.a + f.a + g.a + h.a AS OrderId ,a.a * 10 AS Price ,CONCAT(a.a, N' ', b.a, N' ', c.a, N' ', d.a, N' ', e.a, N' ', f.a, N' ', g.a, N' ', h.a) AS ProductName INTO Table_with_5M_rows FROM a, a AS b, a AS c, a AS d, a AS e, a AS f, a AS g, a AS h;
（3）使用 sqlcmd 連接到數據庫并運行 SQL 腳本以創建包含 5 萬行的表。這可能需要幾分鐘才能運行。
sqlcmd -S 127.0.0.1 -U sa -P <你的> -d SampleDB -i ./CreateSampleTable.sql
3.2、創建一個 Go 應用程序，用于查詢此表并測量所花費的時間
（1）在項目文件夾中，初始化 Go 依賴項。
go get github.com/denisenkom/go-mssqldb go install github.com/denisenkom/go-mssqldb
（2）在您的文件夾中創建一個名為 columnstore.go 的文件。
package main
import (
_ "github.com/denisenkom/go-mssqldb"
"database/sql"
"context"
"log"
"fmt"
"time"
)
var server = "localhost"
var port = 1433
var user = "sa"
var password = "你的"
var database = "SampleDB"
var db *sql.DB
// Delete an employee from database
func ExecuteAggregateStatement(db *sql.DB) {
ctx := context.Background()
// Ping database to see if it's still alive.
// Important for handling network issues and long queries.
err := db.PingContext(ctx)
if err != nil {
log.Fatal("Error pinging database: " + err.Error())
}
var result string
// Execute long non-query to aggregate rows
err = db.QueryRowContext(ctx, "SELECT SUM(Price) as sum FROM Table_with_5M_rows").Scan(&result)
if err != nil {
log.Fatal("Error executing query: " + err.Error())
}
fmt.Printf("Sum: %s\n", result)
}
func main() {
// Connect to database
connString := fmt.Sprintf("server=%s;user id=%s;password=%s;port=%d;database=%s;",
server, user, password, port, database)
var err error
// Create connection pool
db, err = sql.Open("sqlserver", connString)
if err != nil {
log.Fatal("Open connection failed:", err.Error())
}
fmt.Printf("Connected!\n")
defer db.Close()
t1 := time.Now()
fmt.Printf("Start time: %s\n", t1)
ExecuteAggregateStatement(db)
t2 := time.Since(t1)
fmt.Printf("The query took: %s\n", t2)
}
3.3、測量運行查詢所需的時間
從終端運行 Go 應用。
go run columnstore.go
執行結果：
Connected!
Start time: 2023-04-02 15:33:50.0340976 -0700 PDT
Sum: 50000000
The query took: 601.7463ms
3.4、使用 SQLCMD 向表中添加列存儲索引
運行以下命令以在表上創建列存儲索引：
sqlcmd -S localhost -U sa -P <你的> -d SampleDB -Q "CREATE CLUSTERED COLUMNSTORE INDEX Columnstoreindex ON Table_with_5M_rows;"
3.5、重新運行 columnstore.go 腳本，并注意這次完成查詢所花費的時間
go run columnstore.go
Connected!
Start time: 2017-06-05 16:35:02.5409285 -0700 PDT
Sum: 50000000
The query took: 86.9826ms
總結
使用列存儲索引使 Go 應用更快。
參考文檔。
原文地址：http://www.zmynmublwnt.cn/uploads/allimg/xkijul50cys
