使用 to_datetime pandas 方法时 object 是不可分割的错误

Question

I am learning to use pandas library of python. I was trying to use to_datetime method and I have this error 'ValueError: cannot assemble the datetimes: 'int' object is unsliceable'.

Here my code:

import numpy as np
import pandas as pd
from urllib import request
request.urlretrieve ("https://raw.githubusercontent.com/jakevdp/data-CDCbirths/master/births.csv","births.csv")
births= pd.read_csv('births.csv')
births=births.groupby(['year','month','day'],as_index=False).agg('sum')
dates=births.drop(['births'],axis=1)
dates=dates.astype('int')
dates.head()

when I run that the dates df printed seems all right:

year    month   day 
0   1969    1   1  
1   1969    1   2 
2   1969    1   3   
3   1969    1   4 
4   1969    1   5

Then I run:

pd.to_datetime(dates)

and I got the said error. What can it be?

Thanks for insights.

Answer 1

Just pass errors parameter in pd.to_datetime() method and set that equal to 'coerce'

dates=pd.to_datetime(dates,errors='coerce')

Note:- If the above method is creating any problem then use the method below:-

use join() method and for loop :-

datetime=[]
for x in range(0,len(dates)):
    datetime.append(' '.join(['-'.join([str(dates.loc[x,'year']),str(dates.loc[x,'month']),str(dates.loc[x,'day'])])]))

Then:-

dates['date']=datetime

Finally use to_datetime() method:-

dates['date']=pd.to_datetime(dates['date'],errors='coerce')
dates=dates.drop(columns=['year','month','day'])

Output of dates :-

         date
0       1969-01-01
1       1969-01-02
2       1969-01-03
3       1969-01-04
4       1969-01-05
...     ...
7562    1988-12-27
7563    1988-12-28
7564    1988-12-29
7565    1988-12-30
7566    1988-12-31

Answer 2

As pointed out by @serge some days have wrong values making the conversion fails. Once data are filtered it's all fine.

Best