replace with multiple conditions not updating in pandas-CodePudding

I am trying to replace a value based on the row index, and for only certain columns in a dataframe.

for columns b and c, I want to replace the value 1 with np.nan, for rows 1, 2 and 3

df = pd.DataFrame(data={'a': ['"dog", "cat"', '"dog"', '"mouse"', '"mouse", "cat", "bird"', '"circle", "square"', '"circle"', '"triangle", "square"', '"circle"'],
                        'b': [1,1,3,4,5,1,2,3],
                        'c': [3,4,1,3,2,1,0,0],
                        'd': ['a','a','b','c','b','c','d','e'],
                        'id': ['group1','group1','group1','group1', 'group2','group2','group2','group2']})

I am using the following line but its not updating in place, and if I try assigning it, returns only the subset of amended rows, rather than an update version of the original dataframe.

df[df.index.isin([1,2,3])][['b','c']].replace(1, np.nan, inplace=True)

CodePudding user response：

You could do it like this:

df.loc[1:3, ['b', 'c']] = df.loc[1:3, ['b', 'c']].replace(1, np.nan)

Output:

>>> df
                        a    b    c  d      id
0            "dog", "cat"  1.0  3.0  a  group1
1                   "dog"  NaN  4.0  a  group1
2                 "mouse"  3.0  NaN  b  group1
3  "mouse", "cat", "bird"  4.0  3.0  c  group1
4      "circle", "square"  5.0  2.0  b  group2
5                "circle"  1.0  1.0  c  group2
6    "triangle", "square"  2.0  0.0  d  group2
7                "circle"  3.0  0.0  e  group2

A more dynamic version:

cols = ['b', 'c']
rows = slice(1, 3)  # or [1, 2, 3] if you want
df.loc[rows, cols] = df.loc[rows, cols].replace(1, np.nan)