需求是这样的:

有大量名字以序列号排序 内容如下面内容的文本文件

姓名:aaaa

性别:bbb

年龄:ccc

籍贯:ddd

冒号和内容中间有数量不等的空格,如何能批量导入这些文件进入一个数据库，比如access

形如下面的形式

姓名 性别 年龄 籍贯

aaaa1 bbb1 ccc1 ddd1

aaaa2 bbb2 ccc2 ddd2

其中某项空缺的时候，比如有一个文本文件里年龄空缺，怎么保证列表还能正确

下面是测试用的数据文件:外链网址已屏蔽dhxy.info/upload/CS/1134036281_1.rar方案是生成一个csv文件,然后直接导入access,下面是Python实现:

# -*- coding: cp936 -*-

#convert txt data file into csv file

import sys

import os.path

def output(cur):

s = ""

keys = ["名字","性别","年龄","籍贯"]

for key in keys:

if(cur.has_key(key)):

s += cur[key]

s += ","

s+="\n"

return s

def convertFile(filename):

try:

f = file(filename,'r')

fout = file("result.csv",'a')

#fout.writelines("name,sex,age,birthplace\n")

cur = {}

s=f.readline()

tokens = ["名字","性别","年龄","籍贯"]

while len(s) > 0:

lv = s.split(":",2)

if(len(lv)!=2):

s=f.readline()

continue

token = lv[0].lower()

value = lv[1]

if cur.has_key(token) or len(cur)==len(tokens):

result = output(cur)

fout.writelines(result)

cur = {}

if token in tokens:

cur[token] = value.lstrip().rstrip("\n")

else:

print "error line"

s=f.readline()

if len(cur) > 0:

result = output(cur)

fout.writelines(result)

fout.close();

f.close()

except IOError:

print "data file not found!"

#main program

if len(sys.argv) != 2:

print "Usage:python convert.py directory"

sys.exit()

dirname = sys.argv[1]

try:

entryList = os.listdir(dirname)

#remove the result file first and write the header

fout = file("result.csv",'w+')

fout.writelines("名字,性别,年龄,籍贯\n")

fout.close()

for entry in entryList:

if os.path.isdir(dirname + "\\" + entry):

pass

else:

convertFile(dirname + "\\" + entry)

except WindowsError:

print "Dir not found!" + dirname

测试了一下，发现python干这个效率还不赖，6个数据文件，每个5M，P4 2.8G,大概30s跑完。下面是我的同学实现的C代码(跟上一个例子里同一个人:))

#include#include#include#include#defineOUT_FILE  "result.csv"#defineRECORD_MAX_LEN  256struct_record{charname[RECORD_MAX_LEN];charsex[RECORD_MAX_LEN];charage[RECORD_MAX_LEN];charbirth[RECORD_MAX_LEN];intname_len;intsex_len;intage_len;intbirth_len;

};struct_record record;staticchar*token[]={"名字:","性别:","年龄:","籍贯:"};staticinlineintout_record(FILE*out){//fprintf(stdout, "%s,%s,%s,%s ", record.name, record.sex, record.age, record.birth);

fwrite(record.name,1, record.name_len,out);

fwrite(",",1,1,out);

fwrite(record.sex,1, record.sex_len,out);

fwrite(",",1,1,out);

fwrite(record.age,1, record.age_len,out);

fwrite(",",1,1,out);

fwrite(record.birth,1, record.birth_len,out);

fwrite("",1,1,out);

record.name[0]='

测试结果,比Python快11倍....

所以,Python的速度的确比C要慢得多,这些开销主要来自他的动态类型. 在上一个例子中,体现不明显,

性能就不相上下了.