Is there a way in Pandas to fill down the previous value with condition?

Question

I have a table as below and want to fill down the Stage of the same category based on the condition if Stage = "Delivered" then fill down "Delivered" to all the next rows else if Stage = "Paid" then fill down "Paid" to all the next rows

Category	Date	Stage
A	2021-11-01	Ordered
A	2021-12-01	Paid
A	2022-01-01
B	2021-08-01
B	2021-09-01	Ordered
B	2021-10-01	Paid
B	2021-11-01	Ordered
B	2021-12-01	Delivered

The result should look like:

Category	Date	Stage
A	2021-11-01	Ordered
A	2021-12-01	Paid
A	2022-01-01	Paid
B	2021-08-01
B	2021-09-01	Ordered
B	2021-10-01	Paid
B	2021-11-01	Paid
B	2021-12-01	Delivered

Could anyone help? I would really appreciate it!

If condition means per groups then need [dupe](https://stackoverflow.com/questions/38597253/pandas-groupby-forward-fill-with-datetime-index) — jezrael, Jan 12 '22 at 09:08

score 2 · Accepted Answer · answered Jan 12 '22 at 09:58

You can use mask and combine_first:

Assuming your dataframe is already sorted by Date column.

df['Stage'] = df['Stage'].mask(~df['Stage'].isin(['Paid', 'Delivered'])) \
                         .groupby(df['Category']).ffill() \
                         .combine_first(df['Stage'])
print(df)

# Output
  Category        Date      Stage
0        A  2021-11-01    Ordered
1        A  2021-12-01       Paid
2        A  2022-01-01       Paid
3        B  2021-08-01           
4        B  2021-09-01    Ordered
5        B  2021-10-01       Paid
6        B  2021-11-01       Paid
7        B  2021-12-01  Delivered

Is there a way in Pandas to fill down the previous value with condition?

1 Answers1