this is my CODE
import pandas as pd
data1 = {'A': [1, 1, 2, 3, 4, 4, 2, 1], 'B': [2, 2, 3, 4, 4, 4, 1, 2]}
df1 = pd.DataFrame(data1)
df1['OUTPUT'] = df1.duplicated(subset=['A', 'B'], keep=False).map({True: 'OK', False: 'NG'})
print(df1)
and the output is this,but is not i want
i want the output is
update: Sorry, this was my mistake. The values in A and B are strings, not numbers.
If need compare values without order use sorted or frozenset in DataFrame.apply:
df1['OUTPUT'] = (df1[['A', 'B']].apply(sorted, axis=1).duplicated(keep=False)
.map({True: 'OK', False: 'NG'}))
print(df1)
A B OUTPUT
0 1 2 OK
1 1 2 OK
2 2 3 NG
3 3 4 NG
4 4 4 OK
5 4 4 OK
6 2 1 OK
7 1 2 OK
Or use np.sort:
df1['OUTPUT'] = (pd.DataFrame(np.sort(df1[['A', 'B']], axis=1),
index=df1.index).duplicated(keep=False)
.map({True: 'OK', False: 'NG'}))
print(df1)
A B OUTPUT
0 1 2 OK
1 1 2 OK
2 2 3 NG
3 3 4 NG
4 4 4 OK
5 4 4 OK
6 2 1 OK
7 1 2 OK
If need compare next with previous rows without ordering compare Series.shifted values in helper Series:
s = df1[['A', 'B']].apply(sorted, axis=1)
df1['OUTPUT'] = (s.eq(s.shift()) | s.eq(s.shift(-1))).map({True: 'OK', False: 'NG'})
print(df1)
A B OUTPUT
0 1 2 OK
1 1 2 OK
2 2 3 NG
3 3 4 NG
4 4 4 OK
5 4 4 OK
6 2 1 OK
7 1 2 OK
