How can I extract only the date from a timestamp?

user_11138260 0
Viewed 1244

I have dataframe just like that:

df = pd.DataFrame({
    'timestamp': ['2017-06-04 07:59:42', '2017-06-04 07:59:42',
                  '2017-06-04 07:59:42', '2017-06-04 07:59:42',
                  '2017-06-04 07:59:42'],
    'municipality_id': [9, 8, 4, 0, 7],
    'usage': [454, 556, 1090, 204, 718],
    'total_capacity': [1332, 2947, 3893, 2813, 2019]
})
timestamp municipality_id usage total_capacity
2017-06-04 07:59:42 9 454 1332
2017-06-04 07:59:42 8 556 2947
2017-06-04 07:59:42 4 1090 3893
2017-06-04 07:59:42 0 204 2813
2017-06-04 07:59:42 7 718 2019

So how can I get the first ten characters of every row of first column like:

0    2017-06-04
1    2017-06-04
2    2017-06-04
3    2017-06-04
4    2017-06-04
pandas dataframe python
2 Answers

From your csv file, there are multiple methods to extract the date

Method 1: parse_dates from read_csv:

df = pd.read_csv('data.csv', parse_dates=['timestamp'])
df['date'] = df['timestamp'].dt.date

Method 2: to_datetime:

df = pd.read_csv('data.csv')
df['date'] = pd.to_datetime(df['timestamp']).dt.date

Method 3: str[]

df = pd.read_csv('data.csv')
df['date'] = df['timestamp'].str[:10]

The output of each method is:

>>> df['date']
1        2017-06-04
2        2017-06-04
3        2017-06-04
4        2017-06-04
            ...    
13065    2017-08-19
13066    2017-08-19
13067    2017-08-19
13068    2017-08-19
13069    2017-08-19
Name: date, Length: 13070, dtype: object
user_152399510

Method 4: str.split()

df = pd.read_csv('data.csv') 
df['date'] = df['timestamp'].str.split().str[0]

split returns a list of the datetime string separated by spaces

df['timestamp'].str.split()
0    [2017-06-04, 07:59:42]
1    [2017-06-04, 07:59:42]
2    [2017-06-04, 07:59:42]
3    [2017-06-04, 07:59:42]
4    [2017-06-04, 07:59:42]
Name: timestamp, dtype: object

and then you can get the first element of the list by

df['timestamp'].str.split().str[0]
0    2017-06-04
1    2017-06-04
2    2017-06-04
3    2017-06-04
4    2017-06-04
Name: timestamp, dtype: object
user_122877430
Related