转换 WHERE 子句中的多个 'old format' 联接

Question

我在 SQL 服务器中有以下代码结构。它在 WHERE 子句中使用多个旧格式 *= 连接？

SELECT ....
FROM (select ...) L1,
     (select ...) L2,
     (select ...) SE,
     (select ...) EX,
     (select ...) SE,
     (select ...) Y1,
     (select ...) Y2,
     (select ...) Y3,
     (select ...) Y4,
     (select ...) Y5
WHERE SE.FV = EX.FV
    and SE.FV *= Y1.FV
    and SE.FV *= Y2.FV
    and SE.FV *= Y3.FV
    and SE.FV *= Y4.FV
    and SE.FV *= Y5.FV
    and SE.L1 *= L1.FV
    and SE.L2 *= L2.FV
GROUP BY L1.FV, L1.Descr, L2.FV, L2.Descr
ORDER BY L1.FV, L1.Descr, L2.FV, L2.Descr

在实际代码中，select 是动态生成的大杂烩的母亲 sql。但本质上，结构如我上面所示。

上述模板的结构应该是什么，以便它使用正确的 LEFT OUTER JOIN 语法，而不改变查询的逻辑？

这样的东西可能行得通吗？

SELECT L1.FV, L1.Descr, L2.FV, L2.Descr, Y1.B, Y1.A, Y2.B, Y2.A, ...
FROM (select ...) SE
     LEFT OUTER JOIN (select ...) Y1 ON SE.FV = Y1.FV
        AND LEFT OUTER JOIN (select ...) Y2 ON SE.FV = Y2.FV
        AND LEFT OUTER JOIN (select ...) Y3 ON SE.FV = Y3.FV
        AND LEFT OUTER JOIN (select ...) Y4 ON SE.FV = Y4.FV
        AND LEFT OUTER JOIN (select ...) Y5 ON SE.FV = Y5.FV
        AND LEFT OUTER JOIN (select ...) L1 ON SE.L1 = L1.FV
        AND LEFT OUTER JOIN (select ...) L2 ON SE.L2 = L2.FV,
     (select ...) EX
WHERE SE.FV = EX.FV
GROUP BY L1.FV, L1.Descr, L2.FV, L2.Descr
ORDER BY L1.FV, L1.Descr, L2.FV, L2.Descr

Answer 1

我希望，您的查询可以带到下面的结构中。

SELECT      L1.FV, 
            L1.Descr, 
            L2.FV, 
            L2.Descr,
            .. 
FROM                (select ...) L1
JOIN/LEFT JOIN      (select ...) L2 ON L2.Col = ....
JOIN/LEFT JOIN      (select ...) SE ON SE.Col = ....
JOIN/LEFT JOIN      (select ...) EX ON EX.Col = ....
JOIN/LEFT JOIN      (select ...) SE ON SE.Col = ....
JOIN/LEFT JOIN      (select ...) Y1 ON Y1.Col = ....
JOIN/LEFT JOIN      (select ...) Y2 ON Y2.Col = ....
JOIN/LEFT JOIN      (select ...) Y3 ON Y3.Col = ....
JOIN/LEFT JOIN      (select ...) Y4 ON Y4.Col = ....
JOIN/LEFT JOIN      (select ...) Y5 ON Y5.Col = ....
GROUP BY L1.FV, L1.Descr, L2.FV, L2.Descr
ORDER BY L1.FV, L1.Descr, L2.FV, L2.Descr;

Answer 2

您可能希望使用像这样的 CTE 将 "mishmash of dynamically generated sql" 与结构分开：

WITH 
    SE AS (select 1 AS X, 2 AS FV, 3 AS L1, 4 AS L2),
    EX AS (select 1 AS X, 2 AS FV),
    Y1 AS (select 1 AS X, 2 AS FV),
    Y2 AS (select 1 AS X, 2 AS FV),
    Y3 AS (select 1 AS X, 2 AS FV),
    Y4 AS (select 1 AS X, 2 AS FV),
    Y5 AS (select 1 AS X, 2 AS FV),
    L1 AS (select 1 AS X, 3 AS FV, 'y' AS Descr),
    L2 AS (select 1 AS X, 4 AS FV, 'z' AS Descr)
SELECT L1.FV, L1.Descr, L2.FV, L2.Descr
FROM 
     SE
     INNER JOIN EX ON EX.FV = SE.FV
     LEFT JOIN Y1 ON Y1.FV = SE.FV
     LEFT JOIN Y2 ON Y2.FV = SE.FV
     LEFT JOIN Y3 ON Y3.FV = SE.FV
     LEFT JOIN Y4 ON Y4.FV = SE.FV
     LEFT JOIN Y5 ON Y5.FV = SE.FV
     LEFT JOIN L1 ON L1.FV = SE.L1
     LEFT JOIN L2 ON L2.FV = SE.L2
GROUP BY L1.FV, L1.Descr, L2.FV, L2.Descr
ORDER BY L1.FV, L1.Descr, L2.FV, L2.Descr

转换 WHERE 子句中的多个 'old format' 联接

Converting Multiple 'old format' Joins in WHERE Clause

tsql

sql-server

join

database-migration