针对特定场景的 Hive 查询

Question

Suppose my data is like below:

And the output should be as below:

The Conditions are:

1. If the cnsm_id consists Type_cd='P' then I need only that tuple
2. If the cnsm_id with has either type_cd=P or type_cd=A then the record should have max(timestamp) .
3. If the cnsm_id doesn't have the type_cd='P' then it should take the type_cd='A' record with the largest timestamp

Please who knew the answer could help me in solving this query?

Answer 1

SELECT
    CNSM_ID,
    MAX(t),
    TYPE_CD
FROM(
SELECT
    CNSM_ID,
    MAX(TIMESTAMP) AS t, 
    type
FROM
    TBL_NAME
GROUP BY
    CNSM_ID,
    TYPE_CD
ORDER BY t DESC) AS TBL_NAME1
GROUP BY CNSM_ID;