如何在 Oracle 中將多行值轉換為具有逗號分隔值不同 ID 的一行

Question

我正在嘗試使用以下查詢以逗號分隔值返回數據。

select u.employeeid,
       ( select name
         from   roles r
         where  r.id = ur.role_id
       ) userrole
from   users u,
       user_role ur
where  u.id in (select DISTINCT ur.user_id
                from user_role ur)
order by u.employeeid asc;

當前 Output：

EMPLOYEEID USERRROLE
---------------------
1000    ROLE_SUPER_ADMIN
1000    ONBOARDING_CHECKER
1000    ROLE_SUPER_ADMIN
1000    ROLE_APPROVER
1000    ROLE_ONBORDING

現在，當我執行查詢時，我希望表如下：

EMPLOYEEID USERRROLE
--------------------
1000       ROLE_SUPER_ADMIN,ONBOARDING_CHECKER,ROLE_SUPER_ADMIN,ROLE_APPROVER,ROLE_ONBORDING

Answer 1

使用LISTAGG （和 ANSI 連接，而不是傳統的逗號連接和令人困惑的子查詢）。 你可能想要的是：

SELECT u.employeeid,
       LISTAGG( r.name, ',' ) WITHIN GROUP ( ORDER BY r.name ) AS userrole
FROM   users u
       INNER JOIN user_role ur
       ON ( u.id = ur.user_id )
       LEFT OUTER JOIN roles r
       ON ( r.id = ur.role_id )
GROUP BY u.employeeid
ORDER BY u.employeeid asc;

雖然我認為你實際寫的是：

SELECT u.employeeid,
       LISTAGG( r.name, ',' ) WITHIN GROUP ( ORDER BY r.name ) AS userrole
FROM   users u
       CROSS JOIN user_role ur
       LEFT OUTER JOIN roles r
       ON ( r.id = ur.role_id )
WHERE  EXISTS ( SELECT 1
                FROM   user_roles x
                WHERE  u.id = x.user_id )
GROUP BY u.employeeid
ORDER BY u.employeeid asc;

Answer 2

這會做：

SELECT LISTAGG("USERRROLE", ',') WITHIN GROUP (ORDER BY "USERRROLE")
FROM Table1 group by "EMPLOYEEID";

檢查小提琴： http://sqlfiddle.com/#!4/d11c6/4

謝謝！！！！！！！