Hive doesn't support in, exists. How do I write the following query?

问题

I have two tables A and B that both have a column id. I wish to obtain ids from A that are not present in B. The obvious way is:

SELECT id FROM A WHERE id NOT IN (SELECT id FROM B)

Unfortunately, Hive doesn't support in, exists or subqueries. Is there a way to achieve the above using joins?

I thought of the following

SELECT A.id FROM A,B WHERE A.id<>B.id

But it seems like this will return the entirety of A, since there always exists an id in B that is not equal to any id in A.

回答1:

You can do the same with a LEFT OUTER JOIN in Hive:

SELECT A.id
FROM A
LEFT OUTER JOIN B
ON (B.id = A.id)
WHERE B.id IS null

回答2:

Hive seems to support IN, NOT IN, EXIST and NOT EXISTS from 0.13.

select count(*)
from flight a
where not exists(select b.tailnum from plane b where b.tailnum = a.tailnum);

The subqueries in EXIST and NOT EXISTS should have correlated predicates (like b.tailnum = a.tailnum in above sample) For more, refer Hive Wiki > Subqueries in the WHERE Clause

回答3:

Should you ever want to do an IN as so:

SELECT id FROM A WHERE id IN (SELECT id FROM B)

Hive has this covered with a LEFT SEMI JOIN:

SELECT a.key, a.val
FROM a LEFT SEMI JOIN b on (a.key = b.key)

回答4:

if you can using spark sql you can use left anti join.

ex: SELECT A.id FROM A left anti join B on a.id=b.id

来源：https://stackoverflow.com/questions/16804433/hive-doesnt-support-in-exists-how-do-i-write-the-following-query