how to replace dataframe in pandas

Question

sample dataframes are

df1= 
A  B  C
a0 b0 c0
a1 b1 c1
a2 b2 c2
...
an bn cn


df2=
A   B 
x0 y0
x1 y1
x2 y2
...
xm ym

What I want to replace is,

if (ai,bi) = (xj,yj) for some j, then ci = 0

For example,

df3= 
A  B  C
1  4 c0
2  8 c1
2  9 c2
3 12 c3
3 16 c4
4 16 c5


df4=
A   B 
2   8
3  12

and I would like to get below

result=
A  B  C
1  4 c0
2  8  0
2  9 c2
3 12  0
3 16 c4
4 16 c5

I couldn't find how to replace by using np.where . I need your help!

Answer 1

Using pd.concat with fillna and drop_duplicates you can do:

result = (pd.concat([df4,df3],ignore_index=True,sort=False)
            .fillna(0).drop_duplicates(['A','B'],keep='first'))
print (result)
   A   B   C
0  2   8   0
1  3  12   0
0  1   4  c0
2  2   9  c2
4  3  16  c4
5  4  16  c5

you can add .sort_values(['A','B']).reset_index(drop=True) at the end to get the exact same result

Using merge with the parameter how='inner' to get the index to change and then loc these index:

result = df3.copy()
result.loc[df3.reset_index().merge(df4,how='inner')['index'],'C'] = 0
print (result)
   A   B   C
0  1   4  c0
1  2   8   0
2  2   9  c2
3  3  12   0
4  3  16  c4
5  4  16  c5