如何根据一行是否包含另一行中的值组合数据框中的行

Question

I have one dataframe that looks like this, with additonal columns:

ID         Paired_ID      ... 
123_1      123_2
123_2      123_1
456_1      456_2
456_2      456_1
789_1      789_2
789_2      789_1
789_3      789_4
789_4      789_3

What I would like to do is, for a particular ID, take the row where it's Paired_ID is the ID, and combine the two rows into one. I've been trying to use pandas merge (

pd.merge(df, df, left_on="ID", right_on="Paired_ID"

but I'm getting duplicates and can't figure out how to get rid of them.

I would like:

ID_x        Paired_ID_x      ID_y     Paired_ID_y  ...
123_1      123_2             123_2      123_1
456_1      456_2             456_2      456_1
789_1      789_2             789_2      789_1
789_3      789_4             789_4      789_3

Answer 1

The assumption is that every value in ID is in paired_ID.

Compare the ends after the '_' delimiter and create two new dataframes,

Concat the dataframes on the columns axis to get your output.

#this extracts the ends of each value in ID and Paired_ID
A = df.ID.str.split('_').str[-1].astype(int)
B = df.Paired_ID.str.split('_').str[-1].astype(int)

#compare, filter df based on the comparison outcome and add suffixes
filter_1 = df.loc[A.le(B)].reset_index(drop=True).add_suffix('_x')
filter_2 = df.loc[~A.le(B)].reset_index(drop=True).add_suffix('_y')

#concatenate along the columns axis to get outcome
pd.concat([filter_1,filter_2],axis=1)


    ID_x    Paired_ID_x ID_y    Paired_ID_y
0   123_1   123_2       123_2   123_1
1   456_1   456_2       456_2   456_1
2   789_1   789_2       789_2   789_1
3   789_3   789_4       789_4   789_3