Return rows with different values from another dataframe in columns

Question

I am trying to return difference between two data frames but in only some columns.

df1:

df2:

I wrote below to filter difference

df = df.merge (saved, indicator=True, how='left').loc[lambda x: x['_merge']!='both']

And it returned

But I want to return rows that are different in only colA and colB instead of filtering entirely identical rows only, so I can get below dataframe;

what I want:

Answer 1

You can specify parameter on in DataFrame.merge for joined columns:

df = (df.merge(saved, indicator=True, how='left', on=['colA','colB'])
        .loc[lambda x : x['_merge']!='both'])
print (df)
  colA  colB colC_x colC_y     _merge
2    C     3      Y    NaN  left_only
3    D     4      X    NaN  left_only