[FIXED] Replacing a value with its previous value in a column if it is greater

Issue

This is my DataFrame:

import pandas as pd
df = pd.DataFrame(
    {
        'a': [101, 90, 11, 120, 1]
    }
)

And this is the output that I want. I want to create column y:

     a    y
0  101    101.0
1   90    101.0
2   11    90.0
3  120    120.0
4    1    120.0

Basically, values in a are compared with their previous value, and the greater one is selected.

For example for row 1, 90 is compared with 101. 101 is greater so it is selected.

I have done it in this way:

df['x'] = df.a.shift(1)
df['y'] = df[['a', 'x']].max(axis=1)

Is there a cleaner or some kind of built-in way to do it?

Solution

You can use np.fmax to get the maxima without creating an additional column:

df["y"] = np.fmax(df["a"], df["a"].shift(1))

This outputs:

     a      y
0  101  101.0
1   90  101.0
2   11   90.0
3  120  120.0
4    1  120.0

We use np.fmax() to ignore the NaN created when shifting df["a"].

Answered By - BrokenBenchmark