Compare two DataFrames for differences but getting 'Can only compare identically-labeled DataFr-CodePudding

I would like to compare two pandas DataFrames for differences using only the columns that are of dtype bools.

However using the code below I am getting the error below even though the index is for both df's the same on a customer_id level

ValueError: Can only compare identically-labeled DataFrame objects

BOOL_FIELDS = ['is_mobile','is_desktop','is_cancelled','is_existing_customer']

temp_df = pd.DataFrame()
customer_df_2020.set_index('customer_id',inplace=True)
customer_df_2021.set_index('customer_id',inplace=True)

temp_df['sort'] = (customer_df_2020[BOOL_FIELDS] != customer_df_2021[BOOL_FIELDS])

customer_df_2020

customer_id   is_mobile  is_desktop  is_cancelled  is_existing_customer
30293          TRUE      FALSE       FALSE         TRUE
28313          FALSE     TRUE        FALSE         TRUE
19313          FALSE     TRUE        FALSE         TRUE

customer_df_2021

customer_id   is_mobile  is_desktop  is_cancelled  is_existing_customer
30293          FALSE     TRUE        TRUE          FALSE
28313          FALSE     TRUE        FALSE         TRUE
19313          FALSE     TRUE        TRUE          FALSE

CodePudding user response：

Seems some indices are different, is possible extract same in both by Index.intersection:

BOOL_FIELDS = ['is_mobile','is_desktop','is_cancelled','is_existing_customer']

customer_df_2020.set_index('customer_id',inplace=True)
customer_df_2021.set_index('customer_id',inplace=True)

sameidx = customer_df_2020.index.intersection(customer_df_2021.index)

temp_df  = (customer_df_2020.loc[sameidx, BOOL_FIELDS] != 
            customer_df_2021.loc[sameidx, BOOL_FIELDS])