subquery

서브쿼리란

하나의 sql 문에 포함되어있는 또다른 sql 문

주의사항

서브쿼리를 괄호로 감싸서 사용

서브쿼리는 단일행 또는 복수행 비교 연산자와 함께 사용가능

서브쿼리에서는 order by를 사용하지못한다.

ㄴ 예시(이렇게 못씀)

select a.* , (select z.contents from BoardContent z where z.BoardID = a.boardid) as Contents 
from (select BoardID , Subject from BoardContent order by BoardID)a;

서브쿼리 사용가능한 곳

1.select

from
where
having

5.order by

insert 문의 values 절

7 update 문의 set 절

스칼라 서브쿼리

select 절에 사용하는 서브쿼리이다. 스칼라 서브쿼리는 한행 , 한컬럼만을 반환하는 서브쿼리를 말한다. 때문에 여러행이 반환되면 오류가난다.

SELECT T1.C1, (SELECT AVG(T2.C1) FROM T2 T2)
FROM T1 T1;

인라인 뷰

from 절에 사용하는 서브쿼리이다. 기본적으로 from 절에 테이블 명이 오도록 되어있으나 서브쿼리가 from 절에 사용되면 동적으로 생성된 테이블 처럼 사용할수있다. 인라인 뷰는 sql 문이 실행될때만 임시적으로 생성되는 동적인 뷰이기 때문에 데이터베이스에 해당 정보가 저장되지 않는다. 인라인 뷰는 동적으로 조인 방식을 사용하는것과 같다.

SELECT T1.C1, T2.C1, T2.C2
FROM T1 T1,
     (SELECT C1, C2 FROM T2) T2
WHERE T1.C1 = T2.C1;

having 절에서 사용하는 서브쿼리

그룹함수와 함께 사용될때 그룹핑된 결과에 대해 부가적인 조건을 주기위해사용.

SELECT T1.C1, T2.C1, T2.C2
FROM T1 T1, T2 T2
WHERE T1.C1 = T2.C1
GROUP BY T1.C1, T2.C1, T2.C2
HAVING AVG(T1.C1) < (SELECT AVG(C1)
                     FROM T2 );

select a.* , (select z.contents from BoardContent z where z.BoardID = a.BoardID) as Contents from (select BoardID , Subject from BoardContent) a;

서브쿼리는 select 절에서도 사용가능 하며 , from 절에서도 사용이가능함

select a.*,(select  z.contents from BoardContent z where z.BoardID = a.BoardID) as contents  from 
(SELECT  
	  a.BoardID
    ,[Subject]  
    ,cast(Contents as nvarchar) Contents  
    , CONVERT(char(10), a.CreateDate, 120) as CreateDate  
    ,isnull(WorkNames ,'없음') WorkNames 
	,AdminNames

   FROM [dbo].[BoardContent] a Join (SELECT CValue, CType FROM Common WHERE CExplain = 'BoardType' AND CValue = 1) b On a.BoardType = b.CValue  --and isnull(a.GubunPart, '') != '구분없음'
    Join  
   ( 
   SELECT DISTINCT a.BoardID,  STUFF  
                   ((SELECT  ',' + WorkName AS [text()]  
                     FROM     WorkPartBoardRelations b Join WorkPart c ON b.WorkPartID = c.WorkPartID  and b.WorkPartID = Case When 635 <> 0 Then c.WorkPartID Else b.WorkPartID End 
					
                     WHERE  b.BoardID = a.BoardID FOR XML PATH('')), 1, 1, '') AS WorkNames  
      FROM     BoardContent a
   ) dd ON a.BoardID = dd.BoardID 
   left outer Join
   ( 
   SELECT DISTINCT a.BoardID,  STUFF  
                   ((SELECT  ',' + AdminName AS [text()]  
                     FROM     BoardAdminRelations b Join Admin c ON b.AdminID = c.AdminID  --and b.WorkPartID = Case When @WorkPartID <> 0 Then @WorkPartID Else b.WorkPartID End
                     WHERE  b.BoardID = a.BoardID FOR XML PATH('')), 1, 1, '') AS AdminNames  
      FROM     BoardContent a   
   ) ee ON a.BoardID = ee.BoardID 
   left outer join
   WorkPartBoardRelations tt ON a.BoardID = tt.BoardID  and tt.WorkPartID = Case When 635 <> 0 Then 635 Else tt.WorkPartID End
   Join WorkPart wp ON tt.WorkPartID = wp.WorkPartID
   Join (Select BoardID, LanguageID From BoardLanguageRelation Group by BoardID, LanguageID) ff On a.BoardID = ff.BoardID and ff.LanguageID = 2
   Join Admin cc ON a.AdminID = cc.AdminID
   WHERE a.BoardType = 1  
   And Case When 0 = 0 and 0 <> '' then isnull(AdminNames,'') + ' ' + isnull(a.AdminName,'') + ' ' + isnull([Subject],'') + ' ' + isnull(cast(Contents as nvarchar(MAX)),'')   
    When 0 = 1 then [Subject]  
    When 0 = 2 Then AdminNames    
    When 0 = 3 Then a.AdminName     
    When 0 = 4 Then cast(Contents as nvarchar)   
    When 0 = 5 Then GubunPart   
    else '1' End  
   like Case When 0 <= 5 and 0 <> '' then '%' + 0 + '%' else '1' end  
   And case when 1 = 6 Then isnull(AuthorYY, '2001-01-02') Else a.CreateDate End >= '2000-01-01' and case when 1 = 6 Then isnull(AuthorYY, '2001-01-02') Else a.CreateDate End  < DATEADD(DD, 1, '2022-04-12')  
   And a.Status = 1  And (1 != 2 OR cc.AdminKind <> 3 OR cc.AdminType <> 2)
   Group by   a.BoardID  
  ,[Subject]  
  ,cast(Contents as nvarchar)   
  ,a.CreateDate  
  ,WorkNames 
  ,AdminNames) a

oracleSqls

프로젝트하면서 썻었던 sql 복습

SELECT * FROM
	((SELECT DISTINCT 
		TB.COL_NAME AS title
		,a.COL_NAME AS colName
		,a.COL_ID AS COL_ID 
		,c.COL_DEPTH
		,a.ADJUST_DATE AS adjustDate
		,a.RECOMMEND_YN AS recommendYn
	
		FROM 
		KMDB.KMDB_COLLECTIONS a 
		LEFT OUTER JOIN (SELECT COL_ID, COL_DEPTH, UPPER_COL_ID FROM KMDB.KMDB_COLLECTIONS_UPPER) c ON a.COL_ID = c.COL_ID
		LEFT OUTER JOIN(SELECT COL_ID, COL_NAME FROM KMDB.KMDB_COLLECTIONS) TB ON TB.COL_ID = c.UPPER_COL_ID
		)A
		LEFT OUTER JOIN(SELECT 
		   DISTINCT 
		   D.COL_ID

		   ,COUNT(TA.COL_ID) OVER(PARTITION BY TB.TYPE_CLSS) AS DATA_COUNT
		   FROM KMDB.KMDB_COLLECTIONS_REL D 
		   LEFT OUTER JOIN(SELECT DISTINCT COL_ID, COL_NAME FROM KMDB.KMDB_COLLECTIONS) TA ON TA.COL_ID = D.COL_ID  
		   LEFT OUTER JOIN(SELECT DISTINCT COL_ID, REL_ID, TYPE_CLSS FROM KMDB.KMDB_COLLECTIONS_REL) TB ON TB.REL_ID = TA.COL_ID) F ON  A.COL_ID=F.COL_ID)
	WHERE 
		recommendYn = '1'
	AND 
		COL_DEPTH = 2
	AND 
		ROWNUM BETWEEN 1 AND 4
	ORDER BY adjustDate DESC;


=> 각 행마다 컬럼 갯수를 세줌

중복 제거할때(PARTITION BY 사용하여 중복제거 )

	SELECT
	*
	FROM(
	SELECT DISTINCT MIN(p1.person_id) OVER(PARTITION BY p1.person_ID) MIN_PERSONID,
	rownum AS num,
	p1.person_id AS personId,
	p1.title_search,
	p1.kornm AS kornm,
	p1.engnm AS engnm,
	p1.kornm_r AS kornmr,
	SUBSTR(p1.BIRTH_DATE,1,4) AS
	brthYr,
	DECODE(p1.life_yn,'D','Y','N') AS deathYn, SUBSTR(p1.death_date,1,4) AS deathYr,
	p1.debutyear AS debutYr,
	f2.FIELD_NM AS fields,
	p1.FILMOGRAPHY AS filmo,
	p1.PERSONIMG AS Image1,
	CONCAT('https://www.kmdb.or.kr/db/per/', p1.person_id) AS kmdbUrl,
	TO_CHAR(SUBSTR(p3.research_note, 2000, 4000)) AS researchs,
	p5.FIELD_NO AS fieldNo
	FROM
	kmdb.PERSON p1
	left outer JOIN kmdb.FIELD_CODE f2 ON p1.PERSONFIELD = f2.FIELD_NM
	LEFT OUTER JOIN kmdb.PERSON_FIELD p5 ON p1.PERSON_ID = p5.PERSON_ID
	left outer JOIN kmdb.PERSON_RESEARCH p3 ON p1.PERSON_ID = p3.PERSON_ID
	left outer JOIN (SELECT distinct M.TITLE_SEARCH , C.PERSON_ID FROM kmdb.CREDIT_MOVIE C, kmdb.MOVIE_SE M
	WHERE C.MOVIE_ID = M.MOVIE_ID AND C.MOVIE_SEQ = M.MOVIE_SEQ
	) p5 ON p5.person_id = p1.PERSON_ID
	WHERE 1 =1
	)
	WHERE num BETWEEN  1 /**P*/ and  10 /**P*/

zuzu578 / oracleSqls