sql - 大查询 - 加入两个表

Question

我们有两个如下表：

表 A

 Name | Question  | Answer
 -----+-----------+-------
 Bob  | Interest  | art_and_theatre      
 Sue  | Interest  | finances_and_investments
 Sue  | Interest  | art_and_theatre
 Joe  | Interest  | cooking_and_nutrition
 Joe  | Interest  | nutrition_and_drinks
 Joe  | Interest  | eco_life
 Joe  | Interest  | beauty
 Bob  | Interest  | nutrition_and_drinks

表 B（静态）

           Interest                         |   Segment
--------------------------------------------+------------------
art_and_theatre                             |   S1
cooking_and_nutrition, nutrition_and_drinks |   S2 
finances_and_investments                    |   S3
finances_and_investments                    |   S4
technology                                  |   S5
telecommunications                          |   S6
art_and_theatre                             |   S7
art_and_theatre                             |   S8
eco_life, cooking_and_nutrition, beauty     |   S9

预期表

 Name | Question  | Answer
 -----+-----------+-------
 Bob  | Interest  | art_and_theatre      
 Sue  | Interest  | finances_and_investments
 Sue  | Interest  | art_and_theatre
 Joe  | Interest  | cooking_and_nutrition
 Joe  | Interest  | nutrition_and_drinks 
 Bob  | Interest  | nutrition_and_drinks
          (+)
 Bob  | Segment   | S1
 Bob  | Segment   | S7
 Bob  | Segment   | S8
 Sue  | Segment   | S3
 Sue  | Segment   | S4
 Sue  | Segment   | S1
 Sue  | Segment   | S7
 Sue  | Segment   | S8
 Joe  | Segment   | S2
 Joe  | Segment   | S9

如您所见，一个用户可以有多个兴趣，多个兴趣可以属于一个细分。这种 JOIN 在 Big Query 中是否可行？

注意：兴趣列将有一个或多个值。仅当所有值都匹配时，才需要连接段。

Answer 1

以下是 BigQuery 标准 SQL

#standardSQL
select name, question, answer from `project.dataset.tableA`
union all
select distinct name, 'segment' as question, segment as answer
from (
  select answer, segment 
  from `project.dataset.tableB`, 
  unnest(split(interest, ', ')) answer
)
join `project.dataset.tableA`
using(answer)
-- order by question, name, answer    

如果适用于您的问题的样本数据 - 输出是

Answer 2

嗯。. . 我在想union all：join

select a.name, a.question, a.answer
from a
union all
select a.name, 'Segment', b.segment
from a join
     b
     on a.answer = b.interest;

Answer 3

是的，这是可能的，您应该可以使用以下 SQL 来完成

with temp as (
   SELECT a.*, b.*
   FROM TABLEA a
   JOIN TABLEB b
    on a.answer = b.interest
)
SELECT t.Name, t.Question, t.Answer from temp
UNION ALL
SELECT t.Name, 'Segment' as Question, t.Segment as Answer from temp