[sampo]

    import pandas as pd
    import functools
    
    (
        pd.Series(range(10))
        .apply(lambda x: x * 10)
        .where(lambda x: x % 2 == 0)
        .pipe(lambda s: functools.reduce(lambda x, y : x + y, s))
    )

[melx]

Readability isn't the best. Also what you present here is method chaining and not functional pipes.

[gnulinux]

What difference does it make? They're conceptually the same thing. You're mapping immutable data to map/reduce/filter like pure functions to get new data.

[melx]

Improved code readability. The pipe operator was a game changer for me even after 20 years of programming (I know I was probably touching wrong things).

[heyoni]

Is the difference mostly syntactic?

[archarios]

with dot chaining you are restricted to only the functions that the API you're working with provides. With functional pipes you can use whatever functions you want.

[sampo]

In this case, the API provides .pipe, with which you can then use whatever functions you want.