pandas和dataframe

https://blog.csdn.net/weixin_32820767/article/details/82314778  pandas dataframe 读写 与 txt 读写

https://blog.csdn.net/weixin_41888503/article/details/81205203 pandas 追加写入。

https://blog.csdn.net/xiangxianghehe/article/details/72615711 pandas DataFrame数据转为list

python数组或列表删除或读取某一列：np.delete(np数组, 下标, axis=1)

https://blog.csdn.net/manjhOK/article/details/80298448

https://www.jb51.net/article/137636.htm    

 

https://www.jb51.net/article/143477.htm  dataframe 求交并差集合

https://blog.csdn.net/chinacmt/article/details/52192449  dataframe join

https://www.csdn.net/gather_26/MtzaAgxsOTcwLWJsb2cO0O0O.html   list求交并差集合

https://www.cnblogs.com/zenan/p/8616508.html 去重

https://blog.csdn.net/Asher117/article/details/84502952 排序

 

 

https://blog.csdn.net/chang1976272446/article/details/71424961    df数据写入postgresql数据库中。

 

******************************************************************************************************************************************

df处理一列：（如果是行就用row）

import re
pattern=re.compile(r'\d+.*')

def v_trans(osv):
    v=pattern.findall(osv)
    if len(v)!=0:
        osv=v[0]
    return osv

df=profile['os_version'].apply(lambda osv: pd.Series(v_trans(osv)))

profile2=pd.concat([profile['os_version'],df],axis=1)