当我运行以下查询时,我会在结果中排除行,即使我知道它们应该在那里。我把它缩小到第5次加入:
LEFT OUTER JOIN BDNDETL W ON
W.CODE = B.WP
如果我删除该连接以及别名“W”引用的列,则会获得我期望的所有行。 W中有些行基于W.CODE = B.WP在B中没有匹配,但我认为LEFT OUTER条件会捕获这些,但事实并非如此。它似乎只是像INNER JOIN一样对待它。
SELECT
p.p2 AS "Section",
d.d9 AS "Alt WBS",
p.p3 AS "Year",
B.ca1 AS "WBS",
concat(b.ca1,'/',b.ca2,'/',b.wp,b.descrip) AS "Work Package",
B.WP,
w.d9 AS "Cost Category",
B.C3 AS "Provider",
b.c1 AS "TType",
w.d4 AS "WP GLA Cat",
U.COSTSET,
CONVERT(varchar, T.PD_FINISH, 101) period,
SUM(A.ALLOWANCE1 + A.ALLOWANCE2 + A.ALLOWANCE3 + A.ALLOWANCE4 + A.ALLOWANCE5 + A.BASIC + A.BONUS + A.COMMODITY + A.DIRECT + A.ENHANCED + A.NIERS + A.OVERTIME + A.PENSION + A.PILON + A.REDUNDANCY) Value
FROM
TPHASE A JOIN CAWP B ON
B.PROGRAM = A.PROGRAM AND B.CAWPID = A.CAWPID
JOIN MULTPROG X ON
A.PROGRAM = X.SUBPROGRAM
JOIN PROGRAM P ON
A.PROGRAM = P.PROGRAM
JOIN BDNDETL D ON
D.CODE = B.CA1
LEFT OUTER JOIN BDNDETL W ON
W.CODE = B.WP
JOIN COSTDETL U ON
A.CLASS = U.CLASS
JOIN RCUTOFF T ON
A.DF_DATE BETWEEN T.PD_START AND T.PD_FINISH
WHERE
X.MASTER = @master AND
D.BREAKFILE = @wbsfile AND
U.INSTANCE = @costdetlid AND
T.INSTANCE = @cutoffid AND
W.BREAKFILE = @wpbdn
GROUP BY
p.p2,
d.d9,
p.p3,
B.ca1,
concat(b.ca1,'/',b.ca2,'/',b.wp,b.descrip),
B.WP,
w.d9,
B.C3,
b.c1,
w.d4,
U.COSTSET,
T.PD_FINISH
;
答案 0 :(得分:3)
你的where子句中有W.BREAKFILE = @wpbdn
,左连接的空值无法通过此过滤器。
答案 1 :(得分:3)
WHERE
子句中的这个条件:
W.BREAKFILE = @wpbdn
没有匹配时失败。为什么?因为W.BREAKFILE
是NULL
。您应该将其移动到适当的ON
子句:
LEFT OUTER JOIN BDNDETL W ON
W.CODE = B.WP AND W.BREAKFILE = @wpbdn
而且,一条建议。当您混合inner
和left
联接时,我强烈建议您先将所有内部联接放在left join
之后:
FROM TPHASE A JOIN
CAWP B
ON B.PROGRAM = A.PROGRAM AND B.CAWPID = A.CAWPID JOIN
MULTPROG X
ON A.PROGRAM = X.SUBPROGRAM JOIN
PROGRAM P
ON A.PROGRAM = P.PROGRAM JOIN
BDNDETL D
ON D.CODE = B.CA1 JOIN
COSTDETL U
ON A.CLASS = U.CLASS JOIN
RCUTOFF T
ON A.DF_DATE BETWEEN T.PD_START AND T.PD_FINISH LEFT JOIN
BDNDETL W
ON W.CODE = B.WP
这使逻辑更容易理解。我们的想法是内连接进行过滤,然后左连接说“保留所有这些行,如果匹配则添加这些额外的列”。