Replace non-zero values with 1

Question

I have a dataset that looks like this:

id    col1   col2   col3
123   10     0      82
456   0      90     16
987   0      0      0

I'd like to replace all non-zero values in the columns that are not id with 1.

I've tried:

df.col1 = df.where(df.col1 != 0, 1)

But that replaces the non-zeroes with 1s and the zeros with the id.

Then I tried setting a new df to this function:

df2 = df.col1 = df.where(df.col1 != 0, 1)

And that works except it changes the id column to a 1 where it changed the non-zero value to a 1.

Any help troubleshooting the codes that are not working or another approach would be greatly appreciated!

Answer 1

Option 1
astype

df

     col1  col2  col3
id                   
123    10     0    82
456     0    90    16
987     0     0     0

df.astype(bool).astype(int)

     col1  col2  col3
id                   
123     1     0     1
456     0     1     1
987     0     0     0

Or, create a mask with gt:

df.gt(0).astype(int)

     col1  col2  col3
id                   
123     1     0     1
456     0     1     1
987     0     0     0

If id isn't the index, set it first!:

df = df.set_index('id')

---

Option 2
mask

df.mask(df > 0, 1)

     col1  col2  col3
id                   
123     1     0     1
456     0     1     1
987     0     0     0

---

Option 3
df.where (note the difference from your approach)

df.where(df == 0, 1)

     col1  col2  col3
id                   
123     1     0     1
456     0     1     1
987     0     0     0

Similar solution with np.where:

pd.DataFrame(np.where(df > 0, 1, 0), index=df.index, columns=df.columns)

     col1  col2  col3
id                   
123     1     0     1
456     0     1     1
987     0     0     0