How to replace zeros in Pandas Data Frame by negative 1

Question

Replacing negative numbers to zero is straightforward, see How to replace negative numbers in Pandas Data Frame by zero but I want to know how to do the opposite.

I want to replace the 0 here to -1:

But df.replace(0,-1) or df[df == 0] = -1 gives me

It seems like I am able to replace the zeros into positive integers but I cannot replace them with negative integers (i.e. df.replace(0,-3) replaces all my zeros with -253).

Am I missing something here?

Are you sure your columns have type ```int```, and not ```str``` or ```bool```? — iamchoosinganame, May 28 '19 at 18:44

score 3 · Accepted Answer · answered May 28 '19 at 18:45

I believe you have a data type of 8 bit unsigned integer. For that data type, there are no negatives and therefore a -1 overflows(underflows?) to the largest such number.

df = pd.DataFrame([[0, 1], [1, 0]], dtype=np.uint8)

df.replace(0, -1)

     0    1
0  255    1
1    1  255

Where 255 is the largest such number.

np.iinfo(np.uint8).max

255

Instead, set the data type first

df.astype(int).replace(0, -1)

   0  1
0 -1  1
1  1 -1

How to replace zeros in Pandas Data Frame by negative 1

1 Answers1