merge rows with reversed columns-CodePudding

I have dataframe, and I would like to merge the rows that has the same value in reversed columns. An example as below:

Column1 Column2
A       B
B       A
C       D
D       C
E       F

Expected results:

Column1 Column2
A       B
C       D
E       F

As the file has less than 50 lines (though I have 1000 files), I tried some codes use iterrows as followed:

for index, row in df.iterrows():
    output = []
    row_rev = df[(df['Column1'] == row['Column2']) & (df['Column2'] == row['Column1'])]
    row_rev_index = df[(df['Column1'] == row['Column2']) & (df['Column2'] == row['Column1'])].index()
    if row_rev.any():
        print(df[min([index, row_rev_index])])
        output.append(df[min([index, row_rev_index])]) # always print out the first line of the reciprocal lines

but it complains that row_rev_index = df[(df['Column1'] == row['Column2']) & (df['Column2'] == row['Column1'])].index() TypeError: 'Int64Index' object is not callable

CodePudding user response：

Change

row_rev_index = df[(df['Column1'] == row['Column2']) & (df['Column2'] == row['Column1'])].index()

row_rev_index = df[(df['Column1'] == row['Column2']) & (df['Column2'] == row['Column1'])].index

or even shorter

row_rev_index = row_rev.index

CodePudding user response：

This may be what you are looking for:

df = df.groupby(df.apply(lambda x: tuple(set(x)),axis=1)).first()