# coding=utf-8
import os
import re
import string
__author__ = 'peng'
from xml.dom import minidom
import sys
def xmlToTxt(xmlString):
#打开文件 把汉字编号gbk改成utf-8格式
file_xml = open(xmlString, "r").read()
file_xml = file_xml.replace('','')
file_xml = unicode(file_xml, encoding='gbk').encode('utf-8')
doc = minidom.parseString(file_xml)
#解析XML，text和instance的内容
root = doc.documentElement
texts = root.getElementsByTagName("text")
instances = root.getElementsByTagName("instance")
#循环读取instances数据并做数据入库准备
for instance in instances:
print("-------------------------------------------")
print (instance.getAttribute("waveuri").split('/')[1])
text_name = instance.getAttribute("waveuri").split('/')[1]
print("-------------------------------------------")
reload(sys)
sys.setdefaultencoding( "utf-8" )
file_text_name = "e:\\"+text_name+ ".txt"
print(file_text_name.replace(' ',''))
#打开另外一个文件内容，对数据进行填写
file_text = open(file_text_name,"w")
#读取数据并去除标点符号
for text in texts:
print("-------------------------------------------")
print (text.nodeName)
if (len(text.toxml().split(''))==1):
txt = ""
else:
txt =text.toxml().split('')[1].split(' ')[0]
print(type(txt.encode('utf-8')))
#去除标点符号编码，只能一个一个来，如果写循环会出现乱码，原因未知
finalstring = txt.encode('utf-8')
identify = '，'
finalstring = finalstring.replace(identify, "")
identify = '。'
finalstring = finalstring.replace(identify, "")
identify = '！'
finalstring = finalstring.replace(identify, "")
identify = '？'
finalstring = finalstring.replace(identify, "")
print(finalstring)
#最终写入数据
file_text.write(finalstring)
print("-------------------------------------------")
file_text.close()
if ( __name__ == "__main__"):
#文件地址
source = "e:\\" +"文件地址"
#循环读取文件调用方法解析
for root, dirs, files in os.walk( source ):
for OneFileName in files :
print(OneFileName)
xmlToTxt("e:\\"+ "文件地址" +OneFileName) 
python解析xml文档，数据清洗，并入库操作
转载于:https://my.oschina.net/u/2262458/blog/473630