How to replace only the leading zeros with NA?-CodePudding

How can I replace leading zeros with NAs? suppose I have the following example:

import pandas as pd
import numpy as np

df = pd.DataFrame(data={'c1': [0.0, 0.0, 1.0, 0.0], 'c2': [1.0, 1.0, 1.0, 0.0]})

the goal is to have the following result:

 c1    c2
 NA   1.0
 NA   1.0
1.0   1.0
0.0   0.0

but doing the following will not work since it will replace all and not just the leading zeros:

df[np.abs(df) < 1e-50] = np.nan

CodePudding user response：

Similar to mozways' answer, but with assignment and cummin:

df[df.eq(0).cummin()] = pd.NA

Demo:

>>> df
    c1   c2
0  0.0  1.0
1  0.0  1.0
2  1.0  1.0
3  0.0  0.0
>>> df.eq(0).cummin()
      c1     c2
0   True  False
1   True  False
2  False  False
3  False  False
>>> df[df.eq(0).cummin()] = pd.NA
>>> df
    c1   c2
0  NaN  1.0
1  NaN  1.0
2  1.0  1.0
3  0.0  0.0

CodePudding user response：

You could use cummax to fill the non-trailing zeros, and mask those that remain:

df.mask(df.cummax().eq(0))

Or to handle all non zero values with limited precision:

df.mask(df.gt(1e-50).cummax().lt(1e-50))

output:

    c1   c2
0  NaN  1.0
1  NaN  1.0
2  1.0  1.0
3  0.0  0.0

CodePudding user response：

I think you got the sign wrong.

if you switch to

df[np.abs(df) < 1e-50] = np.nan

it should work as intended.

CodePudding user response：

You can also check with ffill

df[df.mask(df==0).ffill().isna()] = np.nan 
df
Out[141]: 
    c1   c2
0  NaN  1.0
1  NaN  1.0
2  1.0  1.0
3  0.0  0.0