add validate='1:1' to merge:

merged_df= pd.merge(df1, df2, on='user_id', how='inner', suffixes=('', '_From_df2'),validate='1:1')