Convert Pandas Column to DateTime With Rare Date Format

Question

I have one column on my dataframe that follows this date format:

17 MAY2016

My code is as follows: df1 =df1.apply(pandas.to_datetime, errors='ignore', format='%d %b%Y')

I also tried: format='%d/%b%Y' format='%d /%b%Y' and still doesn't work. The date column type is still and object. Any ideas? Thanks in advance

jezrael · Accepted Answer · 2017-10-10T11:41:04.323

2

You can use to_datetime only:

df = pd.DataFrame({'date':['17 MAY2016']})

df['date'] = pd.to_datetime(df['date'])
print (df)
        date
0 2016-05-17

If want format parameter:

df['date'] = pd.to_datetime(df['date'], format='%d %b%Y')
print (df)
        date
0 2016-05-17

If some non date values add errors='coerce' for convert them to NaT:

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

EDIT:

For check use dtypes:

print (df.dtypes)
date    datetime64[ns]
dtype: object

edited Oct 10 '17 at 11:41

answered Oct 10 '17 at 11:37

jezrael

first sentence `I have one column` – jezrael Oct 10 '17 at 11:39
Wish I had seen I have one column earlier :( – Bharath M Shetty Oct 10 '17 at 11:59
`df['date'] = pd.to_datetime(df['date'], errors='coerce')` did the trick. But now the date looks like this: "2016-05-17 00:00:00". How could I format it to a: day-month-year date? – User 6683331 Oct 10 '17 at 12:06
I think there are some times. If need dates only need `df['date'] = pd.to_datetime(df['date'], errors='coerce').dt.floor('d')` – jezrael Oct 10 '17 at 12:15

James · Answer 2 · 2017-10-10T11:45:25.690

0

You don't need to use .apply, the to_datetime function natively works on pandas Series objects.

df1['date column'] = pd.to_datetime(df1['date column'], errors='ignore')

edited Oct 10 '17 at 11:45

answered Oct 10 '17 at 11:38

James

2 Answers2