Golang實(shí)現(xiàn)詞頻統(tǒng)計(jì)
來源：程序員人生 發(fā)布時(shí)間：2015-05-14 09:10:01 閱讀次數(shù)：3334次
本例使用golang實(shí)現(xiàn)詞頻統(tǒng)計(jì)。步驟：
（1）從文件中讀取1篇文章。
（2）統(tǒng)計(jì)詞頻，按單詞出現(xiàn)的頻率從大到小進(jìn)行排序。
（3）寫入到文件中。
注：任何非英文字母的符號均認(rèn)為是單詞分隔符（即同等于空格）。
效力：使用本程序統(tǒng)計(jì)1篇150W單詞的文章，大約需要70ms.
1.核心代碼：
package wordtest
import (
"bytes"
"fmt"
"io/ioutil"
"os"
"runtime"
"sort"
"strings"
"time"
)
//簡單的詞頻統(tǒng)計(jì)任務(wù)
func CountTestBase(inputFilePath string, outputFilePath string) {
//時(shí)間開始點(diǎn)
start := time.Now().UnixNano() / 1e6
//讀取文件
fileData, err := ioutil.ReadFile(inputFilePath)
CheckError(err, "read file")
var fileText string = string(fileData)
//根據(jù)CPU核數(shù)新開協(xié)程
newRountineCount := runtime.NumCPU()*2 - 1
runtime.GOMAXPROCS(newRountineCount + 1)
//切分文件
parts := splitFileText(fileText, newRountineCount)
var ch chan map[string]int = make(chan map[string]int, newRountineCount)
for i := 0; i < newRountineCount; i++ {
go countTest(parts[i], ch)
}
//主線程接收數(shù)據(jù)
var totalWordsMap map[string]int = make(map[string]int, 0)
completeCount := 0
for {
receiveData := <-ch
for k, v := range receiveData {
totalWordsMap[strings.ToLower(k)] += v
}
completeCount++
if newRountineCount == completeCount {
break
}
}
//添加進(jìn)slice，并排序
list := make(WordCountBeanList, 0)
for k, v := range totalWordsMap {
list = append(list, NewWordCountBean(k, v))
}
sort.Sort(list)
//時(shí)間結(jié)束點(diǎn)
end := time.Now().UnixNano() / 1e6
fmt.Printf("time consume:%dms
", end-start)
//輸出
wordsCount := list.totalCount()
var data bytes.Buffer
data.WriteString(fmt.Sprintf("程序履行：%dms
", end-start))
data.WriteString(fmt.Sprintf("文章總單詞數(shù)：%d
", wordsCount))
for _, v := range list {
var percent float64 = 100.0 * float64(v.count) / float64(wordsCount)
_, err := data.WriteString(fmt.Sprintf("%s: %d, %3.2f%%
", v.word, v.count, percent))
CheckError(err, "bytes.Buffer, WriteString")
}
err = ioutil.WriteFile(outputFilePath, []byte(data.String()), os.ModePerm)
CheckError(err, "ioutil.WriteFile")
}
func countTest(text string, ch chan map[string]int) {
var wordMap map[string]int = make(map[string]int, 0)
//按字母讀取，除26個(gè)字母（大小寫）以外的所有字符均認(rèn)為是分隔符
startIndex := 0
letterStart := false
for i, v := range text {
if (v >= 65 && v <= 90) || (v >= 97 && v <= 122) {
if !letterStart {
letterStart = true
startIndex = i
}
} else {
if letterStart {
wordMap[text[startIndex:i]]++
letterStart = false
}
}
}
//最后1個(gè)單詞
if letterStart {
wordMap[text[startIndex:]]++
}
ch <- wordMap
}
//將全文分成n段
func splitFileText(fileText string, n int) []string {
length := len(fileText)
parts := make([]string, n)
lastPostion := 0
for i := 0; i < n⑴; i++ {
position := length / n * (i + 1)
for string(fileText[position]) != " " {
position++
}
parts[i] = fileText[lastPostion:position]
lastPostion = position
}
//最后1段
parts[n⑴] = fileText[lastPostion:]
return parts
}
func CheckError(err error, msg string) {
if err != nil {
panic(msg + "," + err.Error())
}
}
2.1個(gè)struct
package wordtest
type WordCountBean struct {
word string
count int
}
func NewWordCountBean(word string, count int) *WordCountBean {
return &WordCountBean{word, count}
}
type WordCountBeanList []*WordCountBean
func (list WordCountBeanList) Len() int {
return len(list)
}
func (list WordCountBeanList) Less(i, j int) bool {
if list[i].count > list[j].count {
return true
} else if list[i].count < list[j].count {
return false
} else {
return list[i].word < list[j].word
}
}
func (list WordCountBeanList) Swap(i, j int) {
var temp *WordCountBean = list[i]
list[i] = list[j]
list[j] = temp
}
func (list WordCountBeanList) totalCount() int {
totalCount := 0
for _, v := range list {
totalCount += v.count
}
return totalCount
}
3.主函數(shù)：
package main
import (
"WordsTest/wordtest"
)
func main() {
inputFilePath := "files/article.txt"
outputFilePath := "files/hanjun-result.txt"
wordtest.CountTestBase(inputFilePath, outputFilePath)
}
生活不易，碼農(nóng)辛苦
如果您覺得本網(wǎng)站對您的學(xué)習(xí)有所幫助,可以手機(jī)掃描二維碼進(jìn)行捐贈
