sql Hive joinkey唯一性判定

Posted 2021-05-13

-- 1. 判断joinkey唯一性
-- 2. 如果不唯一, 判断joinkey在所需维度上的唯一性, 如果唯一, 则没问题(直接对所有所需维度取max即可, 然后根据连接条件group by), 如果不唯一, 需要找上游确认原因

-- joinkey唯一性判断
SELECT
	ad_id AS ad_id,
	ftime AS ftime,
	COUNT( *) as num
FROM
	fdm.fdm_szad_w_ad_info_day
WHERE
	ftime >= '20180426'
	AND platform_type = '京东直投'
	AND media_type = '广点通'
GROUP BY
	ad_id,
	ftime
HAVING
	num > 1
	
-- 在所需列上的唯一性判断
SELECT
	所需维度1,
  所需维度2,
  joinkey,
	COUNT( *) as num
FROM
	fdm.fdm_szad_w_ad_info_day
WHERE
	ftime >= '20180426'
	AND platform_type = '京东直投'
	AND media_type = '广点通'
GROUP BY
	所需维度1,
  所需维度2,
  joinkey
HAVING
	num > 1

-- 若在所需列上唯一, 则直接取max
SELECT
  max(所需维度1),
  max(所需维度2),
  joinkey
FROM
  table_name
GROUP BY
  joinkey