[FIXED] Filling values in dataframe between 2 specific unique values

Issue

i have a following dataframe:

index	Value
1	None
2	A
3	None
4	A
5	B
6	B
7	None
8	A
9	A
10	B

The idea is to fill None in between A and B such that there is no consecutive A or B in the column.

Desired output

index	Value
1	None
2	A
3	None
4	None
5	B
6	None
7	None
8	A
9	None
10	B

This can be done easily with a loop but since I am using pandas and numpy, I hope to avoid the loop approach.

Solution

You can ffill to propagate the non-None forward, then shift, only keep the values that are not identical to this new series with boolean indexing:

df.loc[df['Value'].eq(df['Value'].ffill().shift()), 'Value'] = None

Or with mask:

df['Value'] = df['Value'].mask(df['Value'].eq(df['Value'].ffill().shift()), None)

Output:

   index Value
0      1  None
1      2     A
2      3  None
3      4  None
4      5     B
5      6  None
6      7  None
7      8     A
8      9  None
9     10     B

Intermediates:

   index Value ffill shift     eq
0      1  None  None  None   True
1      2     A     A  None  False
2      3  None     A     A  False
3      4     A     A     A   True
4      5     B     B     A  False
5      6     B     B     B   True
6      7  None     B     B  False
7      8     A     A     B  False
8      9     A     A     A   True
9     10     B     B     A  False

Answered By - mozway