CODEWITHSUNDEEP
pythonpandasdatetimetime-series
Mayur Mahajan
Mayur Mahajan

Reputation: 122

pandas datetime format conversion

My dataset has values in date column as in following format:

date
[datetime.date(2017, 2, 17)]
[datetime.date(2017, 2, 17)]
[datetime.date(2017, 2, 17)]
[datetime.date(2017, 2, 17)]
[datetime.date(2017, 2, 17)]
[datetime.date(2017, 2, 17)]
[datetime.date(2017, 2, 17)]
[datetime.date(2017, 2, 17)]
[datetime.date(2017, 2, 17)]
[datetime.date(2017, 2, 17)]
[datetime.date(2017, 2, 17)]
[datetime.date(2017, 2, 17)]

I want to convert them into datatype of datetime so that I cant perform time series analysis over it.

i have written in following way: but its giving me value error

df_scores['date']=pd.to_datetime(df_scores['date'],format='[datetime.date("%Y, %m, %d")]')

Upvotes: 1

Views: 450

Answers (3)

Abhinav Gupta
Abhinav Gupta

Reputation: 1958

This should work:

df_scores['date']=pd.to_datetime(df_scores['date'],format="[datetime.date(%Y, %m, %d)]")

Upvotes: 0

cs95
cs95

Reputation: 402263

If your column is a column of date objects in a list, this should suffice - 

df.date = pd.to_datetime(df_scores['date'].str[0], errors='coerce')

If you have a column of strings, you could use str.findall to extract the date artifacts, and str.join to join them in a format that to_datetime understands. 

i = df.date.astype(str)\
      .str.findall('\d+')\
      .str.join('/')
df.date = pd.to_datetime(i, errors='coerce')

The astype(str) is required if you have a column of list of strings, rather than a column of strings. You may also need another parameter errors='coerce' if you have malformed data in your column.

df

         date
0  2017-02-17
1  2017-02-17
2  2017-02-17
3  2017-02-17
4  2017-02-17
5  2017-02-17
6  2017-02-17
7  2017-02-17
8  2017-02-17
9  2017-02-17
10 2017-02-17
11 2017-02-17

Upvotes: 1

Ron Kalian
Ron Kalian

Reputation: 3560

if you from datetime import datetime then you should specify dates like this: [datetime(2017, 2, 17)] This will result in entries of type <class 'pandas._libs.tslib.Timestamp'> which you can convert to datetime using .to_datetime()

Upvotes: 0

Related Questions