删除Pandas中列（不是整列）中不需要的内容

Question

如果我只想在时间戳列中显示2016-06-29并删除不需要的部分，例如15:46:43.895000

，该怎么办？

Answer 1

如果是时间戳：

 df['timestamp'] = df['timestamp'].dt.floor('d')

如果字符串：

 df['timestamp'] = df['timestamp'].str.split().str[0]

样品：

df = pd.DataFrame({'timestamp':pd.date_range('2016-06-29 15:46:43.895000', 
                               periods=3, 
                               freq='2000T')})
print (df)
                timestamp
0 2016-06-29 15:46:43.895
1 2016-07-01 01:06:43.895
2 2016-07-02 10:26:43.895

print (type(df.loc[0, 'timestamp']))
<class 'pandas._libs.tslib.Timestamp'>

df['timestamp'] = df['timestamp'].dt.floor('d')
print (df)
   timestamp
0 2016-06-29
1 2016-07-01
2 2016-07-02

df = pd.DataFrame({'timestamp':['2016-06-20 15:46:43.895000', 
                                '2016-06-22 15:46:43.895000',
                                '2016-06-29 15:46:43.895000']})
print (df)
                    timestamp
0  2016-06-20 15:46:43.895000
1  2016-06-22 15:46:43.895000
2  2016-06-29 15:46:43.895000

print (type(df.loc[0, 'timestamp']))
<class 'str'>

df['timestamp'] = df['timestamp'].str.split().str[0]
print (df)
    timestamp
0  2016-06-20
1  2016-06-22
2  2016-06-29

Answer 2

我们可以使用.dt.normalize()方法：

df['timestamp'] = df['timestamp'].dt.normalize()

Answer 3

df['timestamp']=pd.to_datetime(df['timestamp']).dt.strftime('%Y-%m-%d')