TSQL 从 XML 字符串插入记录

Question

我有一个 SQL 查询，它将记录从我传递给它的 XML 字符串插入到表中。该字符串可以包含 1 个或多个节点，因此每个节点都是一条新记录。

这是我的 XML 字符串：

<root>
    <data>
      <segment>
         <trainingEventID>9</trainingEventID>
         <localeID>641</localeID>
         <numOfTeammates>12</numOfTeammates>
         <nonProdHrs>21</nonProdHrs>
         <segmentDate>10/10/2014</segmentDate>
         <trainers>
            <trainer>
               <empID>HUS123</empID>
            </trainer>
            <trainer>
               <empID>Dan123</empID>
            </trainer>
         </trainers>
      </segment>
   </data>
   <data>
      <segment>
         <trainingEventID>9</trainingEventID>
         <localeID>641</localeID>
         <numOfTeammates>12</numOfTeammates>
         <nonProdHrs>21</nonProdHrs>
         <segmentDate>10/25/2014</segmentDate>
         <trainers>
            <trainer>
               <empID>HUS123</empID>
            </trainer>
            <trainer>
               <empID>Dan123</empID>
            </trainer>
         </trainers>
      </segment>
   </data>
</root>

每个segment 都是一个新记录，添加到表中。

现在，我有一个名为trainers 的单独表。对于每个培训师，我还需要在该表中插入一条记录，但它需要有该段的last inserted record id。

这是我的查询：

 INSERT INTO myTable(trainingEventID, localeID, segmentDate, numofTeammates, nonProdHrs)
    SELECT ParamValues.x1.value('trainingEventID[1]', 'INT'),
           ParamValues.x1.value('localeID[1]', 'INT'),
           ParamValues.x1.value('segmentDate[1]', 'DATE'),
           ParamValues.x1.value('numOfTeammates[1]', 'INT'),
           ParamValues.x1.value('nonProdHrs[1]', 'FLOAT')
    FROM   @xml.nodes('/root/data/segment') AS ParamValues(x1);

如何使用从细分插入创建的记录 ID 将培训师插入另一个表？

Answer 1

鉴于问题中此陈述的澄清：

对于每个培训师，我还需要在该表中插入一条记录，但它需要具有该段的最后插入记录 ID。

存在（在问题的 cmets 中找到）：

总共有 4 条记录插入到 trainer 表中，其中 2 条的段 id 为 1，另外 2 条的段 id 为 2。

下面会将此数据插入到具有自动递增 ID 的相关表中。在示例数据中，我稍微改变了 EmpID 值，以更清楚地表明它确实按预期工作。

DECLARE @DocumentID INT, @ImportData XML;

SET @ImportData = N'
<root>
    <data>
      <segment>
         <trainingEventID>9</trainingEventID>
         <localeID>641</localeID>
         <numOfTeammates>12</numOfTeammates>
         <nonProdHrs>21</nonProdHrs>
         <segmentDate>10/10/2014</segmentDate>
         <trainers>
            <trainer>
               <empID>HUS123</empID>
            </trainer>
            <trainer>
               <empID>Dan123</empID>
            </trainer>
         </trainers>
      </segment>
    </data>
    <data>
      <segment>
         <trainingEventID>9</trainingEventID>
         <localeID>641</localeID>
         <numOfTeammates>12</numOfTeammates>
         <nonProdHrs>21</nonProdHrs>
         <segmentDate>10/25/2014</segmentDate>
         <trainers>
            <trainer>
               <empID>HUS1234</empID>
            </trainer>
            <trainer>
               <empID>Dan1234</empID>
            </trainer>
         </trainers>
      </segment>
   </data>
</root>';


DECLARE @Segment TABLE (SegmentId INT NOT NULL IDENTITY(1, 1) PRIMARY KEY,
                        TrainingEventID INT NOT NULL, -- Unique
                        LocaleID INT NOT NULL, -- Unique
                        NumOfTeammates INT,
                        NonProdHrs INT,
                        SegmentDate DATE); -- Unique
-- Ideally create UNIQUE INDEX with the 3 fields noted above
DECLARE @Trainer TABLE (TrainerId INT NOT NULL IDENTITY(1, 1) PRIMARY KEY,
                        SegmentID INT NOT NULL, -- FK to Segment.SegmentID
                        EmpID VARCHAR(50) NOT NULL);

EXEC sp_xml_preparedocument @DocumentID OUTPUT, @ImportData;

-- First pass: extract "Segment" rows
INSERT INTO @Segment
            (TrainingEventID, LocaleID, NumOfTeammates, NonProdHrs, SegmentDate)
   SELECT TrainingEventID, LocaleID, NumOfTeammates, NonProdHrs, SegmentDate
   FROM   OPENXML (@DocumentID, N'/root/data/segment', 2) 
             WITH (TrainingEventID   INT  './trainingEventID/text()', 
                   LocaleID          INT  './localeID/text()',
                   NumOfTeammates    INT  './numOfTeammates/text()',
                   NonProdHrs        INT  './nonProdHrs/text()',
                   SegmentDate       DATE './segmentDate/text()');


-- Second pass: extract "Trainer" rows
INSERT INTO @Trainer (SegmentID, EmpID)
   SELECT seg.SegmentID, trnr.EmpID
   FROM   OPENXML (@DocumentID, N'/root/data/segment/trainers/trainer', 2) 
             WITH (TrainingEventID   INT         '../../trainingEventID/text()',
                   LocaleID          INT         '../../localeID/text()',
                   SegmentDate       DATE        '../../segmentDate/text()',
                   EmpID             VARCHAR(50) './empID/text()') trnr
   INNER JOIN @Segment seg
           ON seg.[TrainingEventID] = trnr.[TrainingEventID]
          AND seg.[LocaleID] = trnr.[LocaleID]
          AND seg.[SegmentDate] = trnr.[SegmentDate];


EXEC sp_xml_removedocument @DocumentID;
-------------------

SELECT * FROM @Segment ORDER BY [SegmentID];
SELECT * FROM @Trainer ORDER BY [SegmentID];

输出：

SegmentId   TrainingEventID   LocaleID   NumOfTeammates   NonProdHrs   SegmentDate
1           9                 641        12               21           2014-10-10
2           9                 641        12               21           2014-10-25

TrainerId   SegmentID   EmpID
1           1           HUS123
2           1           Dan123
3           2           HUS1234
4           2           Dan1234

参考资料：

• OPENXML
• sp_xml_preparedocument
• sp_xml_removedocument