I have two pyspark dataframe with 6 columns and 50000 rows each.

comparison = SparkSQLCompare(
        spark,
        df1,
        df2,
        
        join_columns=['col1', 'col2', 'col3', 'col4', 'col4', 'col5', 'col6'],  
        abs_tol=0,  # Absolute tolerance
        rel_tol=0,  # Relative tolerance
        df1_name='df1',  # Name of the first dataframe
        df2_name='df2'   # Name of the second dataframe
        )

Now When I do comparison if I take any of one column out my all rows match. If I match all 6 keys they I get Number of rows with some compared columns unequal: 50000 otherwise Number of rows with all compared columns equal: 50000. Interestingly If divide columns in set of 3 columns as keys then still they match when I intersect all 6 columns they match- Any idea whats happening here?