Question

我想按列对结果集进行分组，然后将剩余的列合并到json数组中，但是我不确定如何为此汇总结果。

我想要以下输出：

A_ID | Translations
--------------------
 1   | [{"Name": "english_1","LCID": "en-gb"},{"Name": "french_1","LCID": "fr-fr"}]
 2   | [{"Name": "english_2","LCID": "en-gb"},{"Name": "french_2","LCID": "fr-fr"}]

但是我不能在没有聚合器的情况下按A_ID对结果进行分组，所以我得到以下信息

A_ID | Translations
--------------------
 1   | [{"Name": "english_1","LCID": "en-gb"}]
 1   | [{"Name": "french_1","LCID": "fr-fr"}]
 2   | [{"Name": "english_2","LCID": "en-gb"}]
 2   | [{"Name": "french_2","LCID": "fr-fr"}]

这里是一个例子：


DROP TABLE IF EXISTS #tabA;
DROP TABLE IF EXISTS #tabB;
DROP TABLE IF EXISTS #tabC;

go

CREATE TABLE #tabA
(
   Id int not null
);

CREATE TABLE #tabTranslations
(
   translationId int not null,
   Name nvarchar(32) not null,
   aId int not null,   -- Foreign key.
   languageId int not null --Foreign key
);

CREATE TABLE #tabLanguages
(
   languageId int not null,
   LCID nvarchar(32) not null
);

go

INSERT INTO #tabA (Id)
VALUES
    (1),
    (2);

INSERT INTO #tabTranslations (translationId, Name, aId, languageId)
VALUES
    (1, 'english_1', 1, 1),
    (2, 'french_1', 1, 2),
    (3, 'english_2', 2, 1),
    (4, 'french_2', 2, 2);

INSERT INTO #tabLanguages (languageId, LCID)
VALUES
    (1, 'en-gb'),
    (2, 'fr-fr');

go

select
      _a.Id as A_ID,
      (
          select
            _translation.Name,
            _language.LCID
        for json path
      )
from #tabA  as _a
    inner join #tabTranslations as _translation ON _translation.aId = _a.Id
    inner join #tabLanguages as _language ON _language.languageId = _translation.languageId
    -- group by _a.Id ??
;

go

DROP TABLE IF EXISTS #tabA;
DROP TABLE IF EXISTS #tabTranslations;
DROP TABLE IF EXISTS #tabLanguages;

go

替代解决方案：

我知道我可以做到这一点，但是我显然不想对可用的LCID进行硬编码（也许我可以生成sql查询并执行它？但这感觉太复杂了），我也更喜欢一个数组

select
      _a.Id as A_ID,
      (
        SELECT
          MAX(CASE WHEN [LCID] = 'en-gb' THEN [Name] END) 'en-gb',
          MAX(CASE WHEN [LCID] = 'fr-fr' THEN [Name] END) 'fr-fr'
        FOR JSON PATH, WITHOUT_ARRAY_WRAPPER
      ) as b
from #tabA  as _a
    inner join #tabTranslations as _translation ON _translation.aId = _a.Id
    inner join #tabLanguages as _language ON _language.languageId = _translation.languageId
group by _a.Id;

结果：

A_ID | Translations
--------------------
 1   | { "en-Gb": "english_1", "fr-FR": "french_1"}
 2   | { "en-Gb": "english_2", "fr-FR": "french_2"}

Answer 1

如果我对您的理解正确，那么下一种方法可能会有所帮助。使用其他CROSS APPLY运算符和FOR JSON PATH来获得预期的结果：

声明：

SELECT *
FROM #tabA AS t 
CROSS APPLY (
   SELECT _translation.Name AS Name, _language.LCID AS LCID
   FROM #tabA _a
   inner join #tabTranslations as _translation ON _translation.aId = _a.Id
   inner join #tabLanguages as _language ON _language.languageId = _translation.languageId
   WHERE _a.Id = t.Id
   for json path
) _c(Translations)

输出：

Id  Translations
1   [{"Name":"english_1","LCID":"en-gb"},{"Name":"french_1","LCID":"fr-fr"}]
2   [{"Name":"english_2","LCID":"en-gb"},{"Name":"french_2","LCID":"fr-fr"}]

TSQL分组并组合JSON数组中的其余列

1 个答案: