I was spending hours trying to do such a simple thing, I have a dataframe:

  a b c d
0 1 2 3 4
1 5 6 7 8
2 2 3 4 5
3 5 6 7 8
4 1 2 3 4

I have a dictionary:

dic = {'b':6,'d':2}

I would like to do 2 different things :

1. Simply add a row to the df, with this dic using NaN for column 'a' and 'c'
2. Modify a row with a condition a=3, with this dictionary. For this I tried :

df.loc[df['date'] == date, dic.keys()] = dic.values()

Which provide strange results, like values inside a ()

CodePudding user response：

For new row with default RangeIndex use DataFrame.loc:

dic = {'b':6,'d':2}
df.loc[len(df)] = dic

For modify columns by condition working well for me with your solution, for oldier pandas/python version convert keys and values to lists:

dic = {'b':60,'d':20}

df.loc[df['a'] == 1, list(dic.keys())] = list(dic.values())
print (df)
     a   b    c   d
0  1.0  60  3.0  20
1  5.0   6  7.0   8
2  2.0   3  4.0   5
3  5.0   6  7.0   8
4  1.0  60  3.0  20
5  NaN   6  NaN   2

EDIT:

dic = {'b':60,'d':20}
val = 10

m = df['a'] == val
if m.any():
    df.loc[m, list(dic.keys())] = list(dic.values())
else:
    df.loc[len(df)] = {**dic, **{'a':val}}
print (df)

    a   b    c   d
0   1   2  3.0   4
1   5   6  7.0   8
2   2   3  4.0   5
3   5   6  7.0   8
4   1   2  3.0   4
5  10  60  NaN  20