Time until next occurence of value in pandas dataframe

I am trying to find the time (in days) until next occurrence of each specific value. For example, say I have the data below:

    column1   created_at
0   A         2018-09-03
1   B         2018-09-07
2   B         2018-09-08
3   A         2018-09-09
4   B         2018-09-12

The goal is to get time difference for each value in column1 chronologically. At row 3 column1 value is A and its creation date is 2018-09-09. The last creation before that for A is 6 days ago.
Trying to get this:

    column1   created_at  time_diff
0   A         2018-09-03   NaN
1   B         2018-09-07   NaN
2   B         2018-09-08   1
3   A         2018-09-09   6
4   B         2018-09-12   4

>Solution :

Assuming this is the dataframe df

df['created_at'] = pd.to_datetime(df['created_at']) 
df['time_diff'] = df.groupby('column1')['created_at'].diff().dt.days

First row converts is date time, you can skip this if you want.
Second row gives you what you want.

Tick this answer if this solves your problem.