CODEWITHSUNDEEP
pythonpandasjoin
humphrey chan
humphrey chan

Reputation: 91

Pandas Join 2 tables - and add columns value with same column name

Hihi, i wonder if it is possible to join 2 tables - and add columns value with same column name? for example: df1 join df2 by ['Name']

for common columns, would want to combine and add columns to 1 instead of creating 2 columns with Value_1_x & Value_1_y

##expected output like df_final['Value_1'] = (df1['Value_1]' + df2['Value_1'])

for uncommon columns, want to keep the column with original name and value

##expected output like df_final['Value_2'] = df2['Value_2']

since i'm not sure the future df1 & df2 dimension, and it is not always named Value_1 & Value_2, so i wonder if it is possible to process the "adding" while in joining.

df1
Name Category Value_1
Boo   A         1
Boo   B         2
Boo   C         3

df2
Name Type    Value_1  Value_2
Boo   x         0      Null
Boo   y         1      4

Expected Output
Name Category Type    Value_1  Value_2
Boo   A        x         1      Null
Boo   A        y         2      4
Boo   B        x         2      Null
Boo   B        y         3      4
Boo   C        x         3      Null
Boo   C        y         4      4

Upvotes: 2

Views: 434

Answers (1)

jezrael
jezrael

Reputation: 862481

Unfortunately in join or merge is not possible use sum, possible solution is processing after merging all duplicated numeric columns - it means columns ends with _ with columns without _, specified in parameter suffixes:

df = df1.merge(df2, on='Name', suffixes=('_', ''))

#get only numeric df
df1 = df.select_dtypes(np.number)
#filter numeric columns ends with '_'
for c in df1.columns[df1.columns.str.endswith('_')]:
    #pop add column and delete
    df[c.rstrip('_')] += df.pop(c)
print (df)
  Name Category Type  Value_1 Value_2
0  Boo        A    x        1    Null
1  Boo        A    y        2       4
2  Boo        B    x        2    Null
3  Boo        B    y        3       4
4  Boo        C    x        3    Null
5  Boo        C    y        4       4

Upvotes: 1

Related Questions