Merge by several variables on pandas

Question

I want to join two tables by merge, matching 3 common variables in both tables:

Df1

  job         marital month valor
    unemployed   married june   7
    entrepreneur married august 9

Df2

 job         marital month valor1
    unemployed   married june   10
    entrepreneur married august 15

I´m using that:

  df_nuevo=pd.merge(df1, df2, on='month' and 'marital' and 'job', how='left')

And I´d want to obtain:

New df:

   job         marital month valor valor1
    unemployed   married june   7    10
    entrepreneur married august 9    15

But the record are multiplied.

Thanks in advance

Could you display the results you are getting? – Daniel Walker May 15 '20 at 16:51 — Daniel Walker, May 15 '20 at 16:51

score 0 · Accepted Answer · answered May 15 '20 at 17:07

During merges if you don't set the "on" parameter, pandas will use all columns whose names matche between the two dataframes to merge. Something like the following should work:

df_nuevo = pd.merge(df1, df2, how='left')

However this can lead to unseen errors if the columns aren't exact matches (example: ' marital', vs 'marital') so it can be more consistent to specify all 3 columns in a list using the on=['month', 'marital', 'job'] setup that @chris commented.

Merge by several variables on pandas

1 Answers1