# 导入包
import os.path
import pandas as pd
import numpy as np
# 构造一个DataFrame对象
df = pd.DataFrame(np.random.random([5,5]), index=['a','b','c','d','e'], columns=['aa','bb','cc','dd','ee'])
# 生成html文件
fpath = r'C:\Users\Public'
fName = 'pandas_html.html'
# df.to_html(os.path.join(fpath,fName))
# 定义列表
strs = ['<HTML>'] # 'html开始标签
strs.append('<HEAD><TITLE>to_html</TITLE></HEAD>') # html的标题标签
strs.append('<BODY>') # 'body开始标签
strs.append(df.to_html())
strs.append("</BODY></HTML>") # 结束标签
# 把列表元素链接成字符串
html = "".join(strs)
# 字符串写入html文件
file = open(os.path.join(fpath,fName), 'w')
file.write(html)
file.close()
# 读取html文件
# read_html读取的结果是一个DataFrame的list
fullpath = os.path.join(fpath, fName)
print(fullpath)
df = pd.read_html(fullpath)
print(df[0])
# 从网页读取table数据
webpage = 'https://....'
df1 = pd.read_html(webpage)
print(df1[0])
print(df1[1])