使用Go语言实现找出两个大文件中相同的记录

package main

import (
    "bufio"
    "fmt"
    "os"
    "log"
)

// 从文件中读取数据并返回一个map，记录每一行的出现次数
func readFileToSet(filename string) (map[string]bool, error) {
    file, err := os.Open(filename)
    if err != nil {
        return nil, err
    }
    defer file.Close()

    recordSet := make(map[string]bool)
    scanner := bufio.NewScanner(file)
    for scanner.Scan() {
        line := scanner.Text()
        recordSet[line] = true
    }

    if err := scanner.Err(); err != nil {
        return nil, err
    }

    return recordSet, nil
}

// 找出两个文件中相同的记录
func findCommonRecords(file1, file2 string) ([]string, error) {
    // 读取第一个文件到Set
    recordSet, err := readFileToSet(file1)
    if err != nil {
        return nil, err
    }

    // 打开第二个文件并逐行读取
    file, err := os.Open(file2)
    if err != nil {
        return nil, err
    }
    defer file.Close()

    var commonRecords []string
    scanner := bufio.NewScanner(file)
    for scanner.Scan() {
        line := scanner.Text()
        if recordSet[line] {
            commonRecords = append(commonRecords, line)
        }
    }

    if err := scanner.Err(); err != nil {
        return nil, err
    }

    return commonRecords, nil
}

func main() {
    file1 := "file1.txt"
    file2 := "file2.txt"

    commonRecords, err := findCommonRecords(file1, file2)
    if err != nil {
        log.Fatalf("Error finding common records: %v", err)
    }

    fmt.Println("Common Records:")
    for _, record := range commonRecords {
        fmt.Println(record)
    }
}

apple
banana
cherry
grape
orange

pear
banana
grape
watermelon
apple