Qfyilyi

I have a csv file which has column of dates (D/M/Y format) and I want to convert it to days column. I used following approach:

In [1]: import numpy as np

        import pandas as pd

        from pandas import Series, DataFrame



        #f = pd.read_csv(".some_file.csv")

In [2]: f=pd.DataFrame([['1/1/2013', 400, 1000]['2/1/2013', 500, 3000]], columns=['date','flights','distance'])



        f['date']=pd.to_datetime(f['date'], format='%d/%m/%y',errors='ignore')



        f['day']=f['date'].dt.weekday_name

I was expecting the day column. It appeared for the first time and I used that table too. However after clearing kernel I am getting NaT in date and NaN in day. After that days never appeared.

Am I doing anything wrong? If yes, how did day appeared for the first time?

Thanks for reading. Any help will be appreciated.

PS CSV has more than 330K rows. I have to assign day for each date.

The problem is you have incorrectly specified the datetime format and hidden the error by setting errors='ignore'. You need '%Y' for the full year (see Python's strftime directives):

f['date'] = pd.to_datetime(f['date'], format='%d/%m/%Y', errors='coerce')

errors='coerce' will give NaN for non-convertible dates. Or, to raise errors, just omit the errors parameter altogether.

For such problems, it's good practice to look at f.dtypes to see whether your type conversion has succeeded. For example, you should see:

print(f['date'].dtype)

# datetime64[ns]

You should not see:

print(f['date'].dtype)

# object