Data type still shows 'Object' even after using pandas datetime conversion?

Question

I'm working with COVID19 data and have a date column. It was originally as Int64, so I changed it to datetime object using the below function:

us['date'] = pd.to_datetime(us['date'], format='%Y%m%d', errors='ignore')
us['date'] = us['date'].dt.strftime("%m/%d")

When I print out the result, I see all the data in the column as "1/1, 1/4, 2/1 ....", so on. (There is no null value here) I'm trying to use group this by month, so I tried to use another function to get only months as below:

us['date'].dt.strftime("%m")

But I'm getting the error:

AttributeError: Can only use .dt accessor with datetimelike values

I don't understand, since I change the datatype to datetime using pd_datetime, should this column already be datetime, not object? When I checked the data type, it shows 'O'. Why is this so?

Try posting some sample data covering all formats of us['date'] prior any manipulation. — AtanuCSE
– AtanuCSE, Commented Aug 16, 2020 at 20:19
"It was originally as Int64" - then parsing with format='%Y%m%d' doesn't make much sense to me - instead, this should require you to supply a unit, e.g. unit='s' if the datetime data is represented as UNIX time / seconds since the epoch. — FObersteiner
– FObersteiner, Commented Aug 17, 2020 at 6:56

BENY · Accepted Answer · 2020-08-16 20:22:12Z

1

When we do ignore ,it will ignore the whole column if any item can not be changed ,we should do coerce

us['date'] = pd.to_datetime(us['date'], format='%Y%m%d', errors='coerce')
us['date'].dt.strftime('%B')

edited Aug 16, 2020 at 20:22

answered Aug 16, 2020 at 20:05

BENY

324k22 gold badges176 silver badges250 bronze badges

Sign up to request clarification or add additional context in comments.

4 Comments

Sarah Over a year ago

Hi, I still get the exact same attribute error.......

BENY Over a year ago

@sarahkim let us drop na value ~

Sarah Over a year ago

Yes I changed to coerce and tried, same error. pd.to_datetime worked perfectly fine. I changed from 20200101 string --> 2020-01-01 datetime successfully. From here I tried to extract the month 2020-01-01 --> January, is where I'm failing...

BENY Over a year ago

@sarahkim us['date'].dt.strftime('%B') ?

AtanuCSE · Accepted Answer · 2020-08-16 20:21:19Z

0

Lets try this:

us['monthanddate'] = us['date'].dt.strftime("%m/%d")

For only month

us['monthonly'] = us['date'].dt.strftime("%m")

answered Aug 16, 2020 at 20:21

AtanuCSE

8,95016 gold badges80 silver badges115 bronze badges

Collectives™ on Stack Overflow

Data type still shows 'Object' even after using pandas datetime conversion?

2 Answers 2

4 Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

4 Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related