awk, c, java 的IO速度測試

首先用c++獲得的一個簡單但龐大的測試文件：1000萬行的「A B C D」

createFile.cpp 生成的測試文件叫test.in

#include<cstdio>
#define LINENUM 10000000
int main(){
	freopen("test.in","w",stdout);
	for(int i=0;i<LINENUM;++i){
		printf("A\tB\tC\tD\n");
	}
	return 0;
}

awk腳本read.awk

#!/usr/bin/awk -f
BEGIN{
	lines = 0
	start=systime()
}
{
	str = $3
	lines ++
}
END{
	print "line count = " lines
	end=systime()
	print "run time = " end-start "s"
}

c代碼 read.c

#include<stdio.h>
#include<string.h>
#include<time.h>
FILE * fp;
int main(int argc,char * argv[]){
	int i;
	for(i=0;i<argc;++i){
		printf("%s\n",argv[i]);
	}
	if(argc > 1){
		fp = fopen(argv[1],"r");
		if(fp == NULL){
			printf("can't open file\n");
			return 0;
		}
		char str[20];
		char str2[20];
		long lines = 0;
		long start = clock();
		while(fgets(str,16,fp)!=NULL){
			++lines;
			strcpy(str2,str);
		}
		long end = clock();
		printf("lines count = %ld\n",lines);
		printf("time cost = %lf\n",1.0*(end-start)/CLOCKS_PER_SEC);
		fclose(fp);
	}
	return 0;
}

java代碼 read.java

import java.io.*;
import java.util.*;
public class read{
	public static void main(String args[]){
		for(int i=0;i<args.length;++i)	System.out.println(args[i]);
		if(args.length > 0){
			Scanner in;
			try{
				in = new Scanner(new FileInputStream(new File(args[0])));
			}catch(Exception e){
				System.out.println("exception occurs");
				in = new Scanner(System.in);
			}
			String s,s2;
			int lines = 0;
			long start = System.currentTimeMillis();
			while(in.hasNext()){
				s = in.nextLine();
				++ lines;
				s2 = s;
			}
			long end = System.currentTimeMillis();
			System.out.println("lines count = "+lines);
			System.out.println("time cost = "+1.0*(end-start)/1000+"s");
		}
	}
}

測試腳本 testScript.sh

#!/bin/sh
echo "test file is creating"
g++ createFile.cpp -o createFile
./createFile
echo "test file created\n\nnow test awk script"
./read.awk test.in
echo "awk script test finish\n\nnow test c programme language"
gcc read.c -o read
./read test.in
echo "c test finish\n\nnow test Java"
javac read.java
java read test.in
echo "java test finish\n"

測試結果：

test file is creating

test file created

 

now test awk script

line count = 10000000

run time = 4s

awk script test finish

 

now test c programme language

./read

test.in

lines count = 10000000

time cost = 0.870000

c test finish

 

now test Java

test.in

lines count = 10000000

time cost = 16.968s

java test finish