generate col of the first index position until the value changes

Question

I have a df of fruits

    fruit
0    apple
1    apple
2    apple
3    banana
4    apple
5    pear

How could I create indexy -- the first index position +1 until the value changes?

    fruit    indexy
0    apple    1
1    apple    1
2    apple    1
3    banana   4
4    apple    5
5    pear     6

Answer 1

Assuming a range index, you could use it, identify the changing points and ffill:

df['indexy'] = (df.index.to_series().add(1)
                  .where(df['fruit'].ne(df['fruit'].shift()))
                  .ffill().astype(int)
               )

Or, independently of the index, with rank:

df['indexy'] = (df['fruit'].ne(df['fruit'].shift()).cumsum()
                .rank(method='min').astype(int)
               )

Output:

    fruit  indexy
0   apple       1
1   apple       1
2   apple       1
3  banana       4
4   apple       5
5    pear       6

Intermediates (first approach):

    fruit  index+1  change  where  ffill  astype(int)
0   apple        1    True    1.0    1.0            1
1   apple        2   False    NaN    1.0            1
2   apple        3   False    NaN    1.0            1
3  banana        4    True    4.0    4.0            4
4   apple        5    True    5.0    5.0            5
5    pear        6    True    6.0    6.0            6

Intermediates (second approach):

    fruit  ne(shift)  cumsum  rank(min)  astype(int)
0   apple       True       1        1.0            1
1   apple      False       1        1.0            1
2   apple      False       1        1.0            1
3  banana       True       2        4.0            4
4   apple       True       3        5.0            5
5    pear       True       4        6.0            6