使用 SQL 根据硬编码数字对 XML 输出进行分组

Question

我有一个以 XML 形式提供输出的查询。

DECLARE @XMLDATA1 XML 
SET @XMLDATA1 = (


SELECT StudentId AS 'StudentID',
StudentName AS 'StudentName',
StudentAge AS 'Age',
StudentYear 'Year'

FROM Student 
ORDER BY StudentId
FOR XML PATH('StudentLine')
)

SELECT ( CAST(@XMLDATA1 AS XML ))
FOR XML PATH ('StudentRecords') , Root('Student')

示例输出如下。

<StudentRecords>
    <Student>
        <StudentLine>
            <StudentID>12</StudentID>
            <StudentName>Kevin</StudentName>
            <Age>10</Age>
            <Year>5</Year>
        </StudentLine>
        <StudentLine>
            <StudentID>13</StudentID>
            <StudentName>Peter</StudentName>
            <Age>10</Age>
            <Year>5</Year>
        </StudentLine>
        <StudentLine>
            <StudentID>14</StudentID>
            <StudentName>Martin</StudentName>
            <Age>10</Age>
            <Year>5</Year>
        </StudentLine>
        <StudentLine>
            <StudentID>15</StudentID>
            <StudentName>Justin</StudentName>
            <Age>10</Age>
            <Year>5</Year>
        </StudentLine>
    </Student>
</StudentRecords>

我想根据 sql 查询中的硬编码数字将这些记录从 <StudentRecords> 标记中拆分出来。例如，如果我想拆分为 3 条记录，则 xml 将如下所示。

<StudentRecords>
    <Student>
        <StudentLine>
            <StudentID>12</StudentID>
            <StudentName>Kevin</StudentName>
            <Age>10</Age>
            <Year>5</Year>
        </StudentLine>
        <StudentLine>
            <StudentID>13</StudentID>
            <StudentName>Peter</StudentName>
            <Age>10</Age>
            <Year>5</Year>
        </StudentLine>
        <StudentLine>
            <StudentID>14</StudentID>
            <StudentName>Martin</StudentName>
            <Age>10</Age>
            <Year>5</Year>
        </StudentLine>
    </Student>
</StudentRecords>
<StudentRecords>
    <Student>
        <StudentLine>
            <StudentID>15</StudentID>
            <StudentName>Justin</StudentName>
            <Age>10</Age>
            <Year>5</Year>
        </StudentLine>
    </Student>
</StudentRecords>

可能有这样的场景，比如有 100000 个<StudentLine> 标签的记录，硬编码的数字可以是 2000。 有人可以帮助我实现这一目标。 谢谢。

Answer 1

这里是如何通过 XQuery 和 FLWOR 表达式来实现它。

• 组中的行数将作为输入参数。
• 我们将使用NTILE()函数根据数量创建组 组中的行数。
• FLWOR 表达式有两个循环：(1) 遍历组，(2) 遍历给定组中的 <StudentLine> 元素。

SQL

-- DDL and sample data population, start
DECLARE @tbl TABLE (StudentID INT PRIMARY KEY, StudentName VARCHAR(30), Age INT);
INSERT INTO @tbl (StudentID, StudentName, Age) VALUES
(1,  'Martha',  10),
(2,  'Boris',  12),
(3,  'Judy',  15),
(12, 'Kevin',  19),
(13, 'Peter',  10),
(14, 'Martin', 18),
(15, 'Justin', 25);
-- DDL and sample data population, end

-- denominator specifies # of rows in a group
DECLARE @NumOfGroups DECIMAL(10,1) = (SELECT COUNT(*) FROM @tbl) / 3.0;
SET @NumOfGroups = CEILING(@NumOfGroups);

-- just to see
SELECT groupId = NTILE(CAST(@NumOfGroups AS INT)) OVER (ORDER BY StudentID)
    , *
FROM @tbl;

SELECT (
    SELECT groupID = NTILE (CAST(@NumOfGroups AS INT)) OVER (ORDER BY StudentID)
        , *
    FROM @tbl
    FOR XML PATH('StudentLine'), TYPE, ROOT('root')
).query('

    for $gr in distinct-values(/root/StudentLine/groupID/text())
    return <StudentRecords>
                <Student>
            {
                for $x in /root/StudentLine[(groupID/text())[1] eq $gr]
                return $x
            }
        </Student>
    </StudentRecords>

');

输出

  <StudentRecords>
    <Student>
      <StudentLine>
        <groupID>1</groupID>
        <StudentID>1</StudentID>
        <StudentName>Martha</StudentName>
        <Age>10</Age>
      </StudentLine>
      <StudentLine>
        <groupID>1</groupID>
        <StudentID>2</StudentID>
        <StudentName>Boris</StudentName>
        <Age>12</Age>
      </StudentLine>
      <StudentLine>
        <groupID>1</groupID>
        <StudentID>3</StudentID>
        <StudentName>Judy</StudentName>
        <Age>15</Age>
      </StudentLine>
    </Student>
  </StudentRecords>
  <StudentRecords>
    <Student>
      <StudentLine>
        <groupID>2</groupID>
        <StudentID>12</StudentID>
        <StudentName>Kevin</StudentName>
        <Age>19</Age>
      </StudentLine>
      <StudentLine>
        <groupID>2</groupID>
        <StudentID>13</StudentID>
        <StudentName>Peter</StudentName>
        <Age>10</Age>
      </StudentLine>
    </Student>
  </StudentRecords>
  <StudentRecords>
    <Student>
      <StudentLine>
        <groupID>3</groupID>
        <StudentID>14</StudentID>
        <StudentName>Martin</StudentName>
        <Age>18</Age>
      </StudentLine>
      <StudentLine>
        <groupID>3</groupID>
        <StudentID>15</StudentID>
        <StudentName>Justin</StudentName>
        <Age>25</Age>
      </StudentLine>
    </Student>
  </StudentRecords>

Answer 2

这对我有效，无需 FLWOR

-- DDL and sample data population, start
DECLARE @tbl TABLE (StudentID INT PRIMARY KEY, StudentName VARCHAR(30), Age INT);
INSERT INTO @tbl (StudentID, StudentName, Age) VALUES
(1,  'Martha',  10),
(2,  'Boris',  12),
(3,  'Judy',  15),
(12, 'Kevin',  19),
(13, 'Peter',  10),
(14, 'Martin', 18),
(15, 'Justin', 25);
-- DDL and sample data population, end

-- denominator specifies # of rows in a group
DECLARE @NumOfGroups DECIMAL(10,1) = (SELECT COUNT(*) FROM @tbl) / 3.0;
SET @NumOfGroups = CEILING(@NumOfGroups);

-- no FLWOR used
WITH c AS   (
            SELECT groupID = NTILE (CAST(@NumOfGroups AS INT)) OVER (ORDER BY StudentID)
                , *
            FROM @tbl
            )
SELECT      (   SELECT      StudentID, StudentName, Age
                FROM        c
                WHERE       c.groupID = g.groupID
                FOR XML     PATH('StudentLine'), TYPE   ) AS Student
FROM        (
            SELECT DISTINCT groupID FROM c
            ) g
FOR XML     PATH('')