I have the data like this
df['Date']=['05.01.2017','05.01.2017']
I tried
df1= pd.to_datetime(df['Date'])
but it turned a bad result
Id like to get the new data like this
result=[05-01-2017,05-01-2017]
Asked
Active
Viewed 3,709 times
1
-
1How is that a bad result? – roganjosh Nov 14 '18 at 21:31
-
@roganjosh: it probably swaps the day and the month. – Willem Van Onsem Nov 14 '18 at 21:32
-
Possibly. But they got it to a datetime. Either need `day_first` argument or pass a format string and `strf` it back out. This has to be a dupe. – roganjosh Nov 14 '18 at 21:33
-
You either deal with a datetime format inherent to pandas or convert back to a string with the new format. Your desired format (without quotation marks) is not possible in python representations of the data. – roganjosh Nov 14 '18 at 21:39
2 Answers
1
You just need to specify the format of the dates in your column. This works for me from your example. Simple.
pd.to_datetime(df['Date'], format = '%d.%m.%Y')
I'm assuming that the numbers in your dates are day.month.year respectively and not month.day.year. If the latter is true then you should use format = '%m.%d.%Y' instead.
James Fulton
- 322
- 2
- 8
-
I'm not so sure. My _guess_ is that they want to convert back to a string because this would not result in a natural representation that's useful for much. I think it's an XY problem. – roganjosh Nov 14 '18 at 21:43
-
I think since the desired `result=[05-01-2017,05-01-2017]` and there are no string marks around the result dates like there was around the input dates, that they want it in datetime format. However, I shall wait for their response. – James Fulton Nov 14 '18 at 21:46
-
Think you're right on that. Can't test atm so I think it relies on clarification from OP on what exactly they intend. – roganjosh Nov 14 '18 at 21:50
-
0
If you know your dates are always going to be strings in the form of MM.DD.YYYY, and you just wand MM-DD-YYYY instead, then you don't have to deal with any datetime conversions. You can just do a substitution on the string:
>>> from datetime import datetime
>>> mydatestrings = ['05.01.2017','05.02.2017','05.03.2017']
>>> newdates = [ d.replace('.','-') for d in mydatestrings ]
>>> print(newdates)
['05-01-2017', '05-02-2017', '05-03-2017']
On the other hand, if you really want to get all of your dates as 'datetime' objects in Python, you can make a new list like this:
>>> datetime_objs = [ datetime.strptime(d, '%m.%d.%Y') for d in mydatestrings ]
Then format it however you need to:
>>> print(datetime_objs[1].strftime('%m-%d-%Y'))
05-02-2017
>>> print(datetime_objs[1].strftime('%b, %d %Y'))
May, 02 2017
Bill M.
- 1,388
- 1
- 8
- 16
-
This doesn't address `pandas` at all, so will be unusable in its current form – roganjosh Nov 14 '18 at 21:47