PL SQL查询以查找与列表中的所有值匹配的标识符列表

Question

Let's assume I have a table that has the following data:

SHELF_ID    PRODUCT_ID

shelf1      product1
shelf1      product2
shelf1      product3
shelf2      product1
shelf2      product2
shelf3      product1

I made a query 'queryA' that returns from another table

PRODUCT_ID

product1
product2

Now I want to use the 'queryA' in another query to identify which shelfs have at least all the products returned in 'queryA'

by looking at the first table you easily realize its shelf1 and shelf2, but how do I make this in PL SQL?

Thank you

Answer 1

I think this query can help you to find what you want :

Basically what did I do : First I did COUNT(DISTINCT ) to product_id for each shelf and then checked if this count equal or greater than queryA product list that's mean the shelf products match. Then I excluded products if they don't match with queryA with using EXISTS() . If you want to see also not matched products you don't need to use that filter.

--DROP TABLE shelves;
CREATE TABLE shelves
(
    SHELF_ID    VARCHAR(100)
    ,PRODUCT_ID VARCHAR(100)
);

INSERT INTO shelves
VALUES 
 ('shelf1','product1')
,('shelf1','product2')
,('shelf1','product3')
,('shelf2','product1')
,('shelf2','product2')
,('shelf3','product1');

--DROP TABLE queryA;

CREATE TABLE queryA
(
    PRODUCT_ID  VARCHAR(100)
);

INSERT INTO queryA VALUES ('product1'),('product2');

SELECT  * 
FROM shelves S
WHERE S.SHELF_ID IN (           
                    SELECT S.SHELF_ID
                            --,COUNT(DISTINCT S.PRODUCT_ID) PRODUCTCOUNT
                    FROM shelves S
                    GROUP BY S.SHELF_ID
                    HAVING COUNT(DISTINCT S.PRODUCT_ID)>=(SELECT COUNT(DISTINCT PRODUCT_ID) 
                                                            FROM queryA Q  )        
                    )
AND EXISTS (SELECT 1 
            FROM queryA Q
            WHERE S.PRODUCT_ID = Q.PRODUCT_ID
            )

Answer 2

You can do this as:

with products as (
      <your query here)
     )
select s.shelf_id
from shelves s join
     products p
     on s.product_id = p.product_id
group by s.shelf_id
having count(distinct s.product_id) = (select count(*) from products);

Basically, this counts the number of matches on each shelf, and makes sure that the count of matches is the total count of products.

If there are no duplicates in shelves , you can use having count(*) = . . .) having count(*) = . . .) having count(*) = . . .) .

Answer 3

You can do this - no aggregation needed. (Other than the SELECT DISTINCT in the top query; it would be MUCH better if you had an additional table, SHELVES , with SHELF_ID as primary key, then the query could be much more efficient.)

select distinct shelf_id
from   shelves s
where  not exists (select product_id
                   from   queryA   -- Your existing query goes here
                   where  (s.shelf_id, product_id) 
                              not in (select shelf_id, product_id
                                      from   shelves
                                     )
                  )
;