hive difference

Form A

id	name	age	city
1	a	18	beijing
2	b	20	beijing

Form B

id	school	city
1	m	beijing
3	n	beijing
5	p	sichuan

Assuming that the partition of the two tables is city, the current requirement is to find the information that exists in table A but does not exist in table B (AB). The filter condition is "id" and city='beijing':

    

select A.id from(
        (select * from  A where A.city = 'beijing') T1
        left outer join
        (select * from B where B.city = 'beijing') T2
        on A.id = B.id and B.id is NULL
    )

The final result of the above statement is: "1" . That is, it realizes the difference set of two tables based on id.