在表日期之后发生的日期条件上联接表

时间:2019-10-29 16:48:32

标签: sql amazon-redshift

我有一个评论表和一个交易表。我想找出客户提交评论后的交易数量。

查看表如下所示:

+-------------+--------------+-------------+----------------+
| customer_id | review_score | review_date | transaction_id |
+-------------+--------------+-------------+----------------+
| 123         | 4            | 2019-01-01  | 894            |
+-------------+--------------+-------------+----------------+
| 123         | 9            | 2019-05-23  | 897            |
+-------------+--------------+-------------+----------------+
| etc         | etc          | etc         | etc            |
+-------------+--------------+-------------+----------------+

交易表如下:

+-------------+------------------+----------------+
| customer_id | transaction_date | transaction_id |
+-------------+------------------+----------------+
| 123         | 2019-10-01       | 901            |
+-------------+------------------+----------------+
| 123         | 2019-12-04       | 903            |
+-------------+------------------+----------------+
| etc         | etc              | etc            |
+-------------+------------------+----------------+

我希望看到以下内容:

+-------------+--------------+-------------+----------------+-------------+------------------+----------------+
| customer_id | review_score | review_date | transaction_id | customer_id | transaction_date | transaction_id |
+-------------+--------------+-------------+----------------+-------------+------------------+----------------+
| 123         | 4            | 2019-01-01  | 894            | null        | null             | null           |
+-------------+--------------+-------------+----------------+-------------+------------------+----------------+
| 123         | 9            | 2019-05-23  | 897            | 123         | 2019-10-01       | 901            |
+-------------+--------------+-------------+----------------+-------------+------------------+----------------+
| 123         | 9            | 2019-05-23  | 897            | 123         | 2019-12-04       | 903            |
+-------------+--------------+-------------+----------------+-------------+------------------+----------------+
| etc         | etc          | etc         | etc            | etc         | etc              | etc            |
+-------------+--------------+-------------+----------------+-------------+------------------+----------------+

交易是在提交最新评论之后进行的。在某些情况下,客户会进行多次购买并提交评论。我想在提交评论后且提交下一个评论之前加入交易。

我的查询:

with review_cte as (
    select transaction_id
    , customer_id
    , review_date
    , lead(review_date, 1) over (partition by customer_id order by review_date) as review_date_lead
    , review_score
    from review
)
select rev.*
    , b.transaction_id
    , b.customer_id
    , b.transaction_date
from review_cte as rev
join booking b
    on b.unique_customer_id = rev.customer_id
    and b.transaction_date > rev.review_date
    and b.transaction_date < rev.review_date_lead
    and b.booking_id <> rev.booking_id
order by rev.customer_id, rev.review_date
    , b.customer_id, b.transaction_date
;

我的查询的问题是它不会返回所有行,也不会多次加入评论交易ID,因为它满足交易日期为>审阅日期的条件。我说依赖是因为我一直在尝试通过一些连接条件和类似的事情进行

1 个答案:

答案 0 :(得分:1)

这应该做您想要的:

select rev.*,
       b.transaction_id, b.customer_id, b.transaction_date
from (select r.*,
             lead(review_date) over (partition by customer_id order by review_date) as next_review_date
      from review r
     ) r left join
     booking b
     on b.unique_customer_id = r.customer_id and
        b.transaction_date > r.review_date and
        (b.transaction_date < r.review_date_lead or r.review_date_lead is null
        ) and
        b.booking_id <> r.booking_id
order by r.customer_id, r.review_date, b.customer_id, b.transaction_date;