Pandas Date Format Does not convert date

I'm using Pandas version 0.12.0 to import a csv file with dates

The dates are in the following format 'SEP2005'

using pandas to read the csv file:

import pandas as pd

DF = pd.read_csv('mydata.csv') 

mydata.head()
Out[40]:
      Date  Quantity
0  APR2002  282.0000
1  APR2002  NaN
2  APR2002  0.0000
3  APR2002  20.2253
4  APR2002  55.6853

I then turn the Date Column to the index using the follow:

mydata.index = pd.to_datetime(mydata.pop('Date'))

Here is what is very strange in the past it has parsed my dates and turned the format into

2002-04-15 which is what I want. Then I would just make sure the days where set the the last day of the month:

mydate.index = mydata.index.to_period('M').to_timestamp('M')

Pandas in the past has done a great job of picking the best date format.

However, When I do this now I'm getting my DataFrame back with the same text "APR2002"

As you would guess the last to_period will not work on that.

I have not change my code and I have not updated Pandas so I'm not sure where this change in coming from.

I'm not sure if I care too much about the why. What I really need help with is how do I format the index column to reflect Year-Month-Day or %Y%m%d as in 2005-04-30

I'm coming from R so any help would be huge!


You could try

 pd.to_datetime(mydata.pop('Date'), format="%b%Y")

but that would expect the date to appear like Apr2002 (note not all caps).

You can specify a datetime format using the format string, and the format string will accept strftime arguments (defined here). There is some pandas documentation on this too.