import xlwt
import openpyxl
import xlrd
def eq(l):
workbook = xlrd.open_workbook('data.xlsx')
table = workbook.sheet_by_index(0)
all_data=[]
url=[]
test=[]
for i in range(0,table.nrows):
u = table.cell(i, 1).value
ip=table.cell(i,0).value
#print(ip)
if u not in l:
print(u,"\t",ip)
def chooseInfo(dir):
l=[]
list = os.listdir(dir)
for i in range(0, len(list)):
path = os.path.join(dir, list[i])
if os.path.isfile(path):
with open(path, encoding="utf-8") as f:
content = f.read()
doc = pq(content) # 解析html 文本
item = doc("h1")
s=((item.eq(2).text())) #获取相应标签文本
#if int(item.eq(1).html()) > 0 or int(item.eq(3).html()) > 0:
l.append(s)
#print("操作全部完成!")
return(l)
a=[]
a=chooseInfo("C:\\Users\\yxb\\Downloads\\新")
eq(a)