Replace values with nan on multiple columns based on condition

Question

My df contains many columns. I want to replace all values only in columns A and B with NaN according to a condition. Also, I want to apply the same condition to another df except on columns C and D. My search so far returns methods that work for a single column.

My attempt so far.
Only on columns A and B:

df[df.loc[:, df.columns['A','B']] < (0.1 * 500)] = np.nan

Except columns A and B:

df[df.loc[:, df.columns != ['A','B']] < (0.1 * 500)] = np.nan

These code doesn't work.

Answer 1

I think you need DataFrame.mask :

df = pd.DataFrame({
         'A':[4,5,4,5,5,4],
         'B':[7,8,9,4,2,3],
         'C':[1,3,5,7,1,0],
         'D':[5,3,6,9,2,4],

}) * 10

c = ['A','B']
df[c] = df[c].mask(df[c] < (0.1 * 500))
print (df)
      A     B   C   D
0   NaN  70.0  10  50
1  50.0  80.0  30  30
2   NaN  90.0  50  60
3  50.0   NaN  70  90
4  50.0   NaN  10  20
5   NaN   NaN   0  40

---

c1 = df.columns.difference(c)
df[c1] = df[c1].mask(df[c1] < (0.1 * 500))

print (df)
      A     B     C     D
0   NaN  70.0   NaN  50.0
1  50.0  80.0   NaN   NaN
2   NaN  90.0  50.0  60.0
3  50.0   NaN  70.0  90.0
4  50.0   NaN   NaN   NaN
5   NaN   NaN   NaN   NaN

Answer 2

# Initalize columns to modify
columns

# Get row indexes where condition is met
indexes=np.where(condition is True)

# Then replace
for col in pest_cols_to_NA:
    tuta_df[col][indexes]=np.nan