pandas pivot_table() 按日期分多列数据的方法

>>> df = pd.Data .from_dict([[\'ip4\',\'20170307\',1],[\'ip4\',\'20170307\',1],[\'ip5\',\'20170307\',1],[\'ip5\',\'20170307\',1],[\'ip6\',\'20170308\',1],[\'ip5\',\'20170308\',1]])

>>> df.columns=[\'type\',\'date\',\'num\']

>>>df

 type   date num
0 ip4 20170307  1
1 ip4 20170307  1
2 ip5 20170307  1
3 ip5 20170307  1
4 ip6 20170308  1
5 ip5 20170308  1

>>> pd.pivot_table(df,values=\'num\',rows=[\'type\'],cols=[\'date\'],aggfunc=np.sum).fillna(0)