How to find indices/rows where a combination of columns are equal to columns another dataframe

I have two dataframes. The first one is with multiple entries of the same values in rows and the other only has one row for each combination of columns. I want to get the indices/rows of the first dataframe where column A, B and C are pairwise equal in the two dataframes and where column D in the first dataframe is equal to n. Column A, B and C are a subset of all columns. I want to specify that only those columns have to be equal.
With those indices I want to change the value of column E (simply using df.loc[indices, 'E']='n'.

The first one is the one with multiple entries for same values. The original dataframe has date stamps so it doesn't have multiple completely identical rows.

   A B C D E
0  a b c y y
1  a b c y y
2  b n m n y
3  b n m n y
4  b n m n y
5  t u j y y
6  t u j y y
7  t u j y y
8  e t y y y
9  e t y y y

The second one only has one row per entry and not column D and E and is a subset of the entries in the first dataframe.

   A B C
0  a b c
1  b n m
2  t u j

In this example the desired output is the list of indices where all conditions are fulfilled.

[2,3,4]

I've tried with merge, while that does give me the right rows, I don't get the right indices to be used to update the first dataframe. I've also tried MultiIndex on the three columns but then I'm missing the filter on column D.

How can I achieve this?

One way to get the list of indexes is to merge the two dataframes, adding the index as a column to the first, and assigning 'n' to the D column in the second so that only rows which match on A, B and C and have 'n' in D will match. Then you can take the index column from the first as your result:

indices = (df1
    .reset_index(names='index')
    .merge(df2.assign(D='n'), on=['A', 'B', 'C', 'D'])['index']
    .to_list())

Output (for your sample data):

[2, 3, 4]

Note that the to_list() is not strictly necessary, the following code will work without it.

You can then use the indices to set E to 'n':

df1.loc[indices, 'E'] = 'n'

Output:

   A  B  C  D  E
0  a  b  c  y  y
1  a  b  c  y  y
2  b  n  m  n  n
3  b  n  m  n  n
4  b  n  m  n  n
5  t  u  j  y  y
6  t  u  j  y  y
7  t  u  j  y  y
8  e  t  y  y  y
9  e  t  y  y  y

Assuming df1 and df2, you can use boolean indexing:

# are the common columns (A/B/C) also in df2?
m1 = df1.merge(df2.assign(mask=True), how='left')['mask'].fillna(False)
# or
import numpy as np
m1 = np.isin(df1[df2.columns], df2).all(1)

# is "D" a "n"?
m2 = df1['D'].eq('n')

# if both conditions are True, update E
df1.loc[m1&m2, 'E'] = 'n'

Output:

    A  B  C  D  E
0   a  b  c  y  y
1   a  b  c  y  y
2   b  n  m  n  n
3   b  n  m  n  n
4   b  n  m  n  n
5   t  u  j  y  y
6   t  u  j  y  y
7   t  u  j  y  y
8   e  t  y  y  y
9   e  t  y  y  y
10  e  t  y  n  y  # not updated

Target indices:

df1.index[m1&m2]
# Index([2, 3, 4], dtype='int64')