How do you change the repr of 0 in Python?

Question

I am building a data pipeline in Pandas, and any given number can be

Explicitly: 0
Missing (np.nan)
Missing, but imputed to have a value of 0

I'd like to replace the imputed 0's in the table with a value that indicates that they have been imputed. The best I could think of was a "0" where the repr is "-"

I tried:

class MyNaNThatIsLikeZero(int):
    def __repr__(self):
        return '-'

myNaN = MyNaNThatIsLikeZero()

But pandas is converting this to just the float 0.0

To clarify, I need this value to behave as 0, so it needs to implement all the dunder methods

You must understand, usually and ideally, **there are no `int` or `float` objects in your dataframe at all**. Just add a column with a categorical variable detailing this distinction — juanpa.arrivillaga, Jan 18 '22 at 15:39

score 0 · Answer 1 · answered Jan 18 '22 at 15:25

0

pandas is converting this to just the float 0.0

You can disable that conversion by passing dtype=object to the Dataframe constructor. See my other answer: https://stackoverflow.com/a/66907030/13622927

answered Jan 18 '22 at 15:25

Expurple

877
6
13

How do you change the __repr__ of 0 in Python?

1 Answers1

How do you change the repr of 0 in Python?