在SQL中,主表应与多列的查找表匹配

问题描述 投票:1回答:1

主表应与给定列的查找表匹配,如果它不匹配且查找表具有任何空值,则它应与Null值匹配(此处为null =任何值)

脚本

CREATE TABLE dbo.MAIN_TABLE (
NAMEID VARCHAR(50) NOT NULL,
COL1 VARCHAR(50) NULL,
COL2 VARCHAR(50) NULL,
COL3 VARCHAR(50) NULL,
COL4 VARCHAR(50) NULL,
CONSTRAINT PK_MAIN_TABLE PRIMARY KEY (NAMEID)
)

CREATE TABLE dbo.LOOKUP_TABLE (
COL1 VARCHAR(50) NULL,
COL2 VARCHAR(50) NULL,
COL3 VARCHAR(50) NULL,
COL4 VARCHAR(50) NULL,
RES_COL VARCHAR(50) NOT NULL
)

TRUNCATE TABLE dbo.LOOKUP_TABLE
INSERT INTO dbo.LOOKUP_TABLE VALUES('AA','BB', 'CC', NULL, 'Rule1')
INSERT INTO dbo.LOOKUP_TABLE VALUES('AA','BB', NULL, NULL, 'Rule2')
INSERT INTO dbo.LOOKUP_TABLE VALUES('AA',NULL, NULL, NULL, 'Rule3')
INSERT INTO dbo.LOOKUP_TABLE VALUES(NULL,NULL, NULL, NULL, 'Rule4')

TRUNCATE TABLE dbo.MAIN_TABLE
INSERT INTO dbo.MAIN_TABLE VALUES('NameId1','AA','BB', 'CC', 'DD' )
INSERT INTO dbo.MAIN_TABLE VALUES('NameId2','AA','BB', 'CC', NULL )
INSERT INTO dbo.MAIN_TABLE VALUES('NameId3','AA','BB', NULL, NULL )
INSERT INTO dbo.MAIN_TABLE VALUES('NameId4','AA', NULL, NULL, NULL )

INSERT INTO dbo.MAIN_TABLE VALUES('NameId5',NULL,'BB', 'CC', 'DD' )
INSERT INTO dbo.MAIN_TABLE VALUES('NameId6',NULL,NULL, 'CC', 'DD' )

预期结果

NAMEID  |  RES_COL
-----------------
NameId1 | Rule1
NameId2 | Rule1
NameId3 | Rule2
NameId4 | Rule3
NameId5 | Rule4
NameId6 | Rule4

我尝试了类似下面的查询,但它返回了包含所有其他值的重复行

SELECT MN.NAMEID, MAP.RES_COL  FROM MAIN_TABLE MN
CROSS APPLY
(
  SELECT * FROM LOOKUP_TABLE LKP
  WHERE  
   ( CHECKSUM(MN.COL1) = CHECKSUM(LKP.COL1) OR LKP.COL1 IS NULL )
  AND
   ( CHECKSUM(MN.COL2) = CHECKSUM(LKP.COL2) OR LKP.COL2 IS NULL)
  AND
   ( CHECKSUM(MN.COL3) = CHECKSUM(LKP.COL3) OR LKP.COL3 IS NULL)
  AND
   ( CHECKSUM(MN.COL4) = CHECKSUM(LKP.COL4) OR LKP.COL4 IS NULL )
) MAP

ORDER BY MN.NAMEID ASC
sql sql-server tsql
1个回答
2
投票

你得到多个匹配因为NULL =匹配任何规则。所以对于任何名称,Rule4肯定会匹配,因为它对所有列都是NULL

如果你只想要最好的比赛,请使用TOP 1ORDER BY

SELECT  *
FROM    dbo.MAIN_TABLE m
        OUTER APPLY
        (
            SELECT  TOP 1 *
            FROM    dbo.LOOKUP_TABLE l
            WHERE   (l.COL1 = m.COL1 OR l.COL1 IS NULL)
            AND     (l.COL2 = m.COL2 OR l.COL2 IS NULL)
            AND     (l.COL3 = m.COL3 OR l.COL3 IS NULL)
            AND     (l.COL4 = m.COL4 OR l.COL4 IS NULL)
            ORDER BY 
                CASE WHEN l.COL1 IS NOT NULL THEN 1 ELSE 2 END
            +   CASE WHEN l.COL2 IS NOT NULL THEN 1 ELSE 2 END
            +   CASE WHEN l.COL3 IS NOT NULL THEN 1 ELSE 2 END
            +   CASE WHEN l.COL4 IS NOT NULL THEN 1 ELSE 2 END
        ) l

你使用CHECKSUM()而不是简单的=的任何特殊原因?

© www.soinside.com 2019 - 2024. All rights reserved.