Question

我正在尝试查询与特定offer_id相关联的消息对。使用子查询：

WITH X AS (
SELECT M.offer_id
FROM quotes_site.messages M
where (regexp_substr (REPLACE(M.body, ' ', ''), '[0][0-9]{8,12}') != '') and (M.body NOT iLIKE '%agreement%' AND M.author_type = 'User') AND
M.created_at BETWEEN '2016-01-03' AND '2016-02-03'
)

然后通过将X与原始表连接，我得到与此id相关联的主题（对于所有可用的author_type）

SELECT MSG.body, MSG.offer_id, MSG.author_type
FROM quotes_site.messages MSG
INNER JOIN X
ON MSG.offer_id = X.offer_id
ORDER BY MSG.offer_id, MSG.created_at

问题是我如何让表MSG包含我在子查询中找到的消息，而且只包含按时间顺序排列在它之前的消息（但是对于同一个offer_id）

Answer 1

您可以将ROW_NUMBER与PARTITION BY子句一起使用，以便枚举属于同一offer_id组的邮件：

WITH M AS (
   SELECT *,
          ROW_NUMBER() OVER (PARTITION BY offer_id 
                             ORDER BY created_at DESC) AS rn
   FROM quotes_site.messages 
) X AS (
   SELECT M.offer_id, M.rn              
   FROM M
   WHERE (regexp_substr (REPLACE(M.body, ' ', ''), '[0][0-9]{8,12}') != '') 
         AND (M.body NOT iLIKE '%agreement%' AND M.author_type = 'User') 
         AND (M.created_at BETWEEN '2016-01-03' AND '2016-02-03')
)
SELECT M.body, M.offer_id, M.author_type
FROM M
INNER JOIN X ON MSG.offer_id = X.offer_id AND M.rn BETWEEN X.rn AND x.rn - 1
ORDER BY M.offer_id, M.created_at

选择数据点及其前面的数据点

1 个答案: