Convert Pandas column object to date-CodePudding

I'm trying to convert all the dates in a pandas column from object to date. The object column is called PRODUCT_CREATION_DATE, and has values like:

2021-12-09 00:00:00-05:00

2021-04-20 00:00:00-07:00

2017-04-25 00:00:00-07:00

I already know I can convert them with pd.to_datetime, but I don't seem to be able to find the correct format, and I keep getting a format error. The problem is the last part, I have no idea what those -05:00 or -07:00 should represent. I actually only care about year, month and day, so if there's a way to get a date with only year, month and day, it would be even better....

This is the code I'm trying to use:

pd.to_datetime(df_info['PRODUCT_CREATION_DATE'], format='%Y-%m-%d %H:%M:%S-%f')

This returns

ValueError: unconverted data remains: :00

CodePudding user response：

The code %f is used for microseconds, whereas in your sample dates, -05:00 refers to the UTC timezone offset. So, the format you need to use with to_datetime will be

%Y-%m-%d %H:%M:%S%z

CodePudding user response：

This will get you only the yyyy-mm-dd format as a datetime you were looking for

df['Date'] = pd.to_datetime(df['Date'].apply(lambda x : x.split(' ')[0]))

CodePudding user response：

If you just need the date, let's just extract the date:

                        col1
0  2021-12-09 00:00:00-05:00
1  2021-04-20 00:00:00-07:00
2  2017-04-25 00:00:00-07:00

df['col1'] = pd.to_datetime(df['col1'].str[:10])
print(df)
print(df.dtypes)

        col1
0 2021-12-09
1 2021-04-20
2 2017-04-25

col1    datetime64[ns]
dtype: object