在 SQL Server 中使用带有 XML 的两个表中的公共字段答案

【问题标题】：Use a common field in two tables with XML in SQL Server在 SQL Server 中使用带有 XML 的两个表中的公共字段
【发布时间】：2021-01-22 15:03:25
【问题描述】：

如何删除#XMLItm表中的冗余？

我的代码：

  BEGIN TRY
    DECLARE @xml XML= '<?xml version="1.0" encoding="iso-8859-1"?>
                <lists>
                    <list list="LIST A" remark="xx">
                        <item id_list="1" code="Y" description="Yes" order="1" enabled="1" />
                        <item id_list="1" code="N" description="No" order="2" enabled="1" />
                    </list>
                    <list list="LIST B" remark="yy">
                        <item id_list="2" code="E" description="Enabled" order="1" enabled="1" />
                        <item id_list="2" code="D" description="Disabled" order="2" active="1" />
                    </list>
                </lists>';
    CREATE TABLE #XMLLst
    (
        [ID_LIST]      INT NOT NULL, 
        [NAME_LIST]  VARCHAR(250) NOT NULL, 
        [REMARKS] VARCHAR(8000)
    );
    INSERT INTO #XMLLst
    (
        ID_LIST, 
        NAME_LIST, 
        REMARKS
    )
           SELECT DISTINCT 
                  ROW_NUMBER() OVER (ORDER BY (SELECT NULL)) AS ID_LIST,
                  Ssn.value('(/list/@list)[1]', 'Varchar(250)') AS NAME_LIST, 
                  Ssn.value('(/list/@remark)[1]', 'varchar(8000)') AS REMARKS
           FROM
           (
               SELECT S.s.query('.') AS Ssn
               FROM @xml.nodes('/lists/list') AS S(s)
           ) AS SSnes;
    WITH CTE_DATOS_ORIGEN
         AS (SELECT DISTINCT 
                    ID_LIST, 
                    NAME_LIST, 
                    REMARKS
             FROM #XMLLst)
         MERGE INTO [sos].[LIST] LIST
         USING CTE_DATA_ORIGIN
         ON LIST.ID_LIST = CTE_DATA_ORIGIN.ID_LIST
             WHEN MATCHED AND CTE_DATA_ORIGIN.NAME_LIST = LIST.NAME_LIST
             THEN UPDATE SET 
                             LIST.NAME_LIST = CTE_DATA_ORIGIN.NAME_LIST, 
                             LIST.REMARKS = CTE_DATA_ORIGIN.REMARKS
             WHEN NOT MATCHED
             THEN
               INSERT(ID_LIST, 
                      NAME_LIST, 
                      REMARKS)
               VALUES
         (
            CTE_DATA_ORIGIN.ID_LIST, 
            CTE_DATA_ORIGIN.NAME_LIST, 
            CTE_DATA_ORIGIN.REMARKS
         );

    CREATE TABLE #XMLItm
    (
        [ID_OPTION_LIST] INT NOT NULL, 
        [ID_LIST]        INT NOT NULL, 
        [CODE]          VARCHAR(50) NOT NULL, 
        [DESCRIPTION]     VARCHAR(1000) NOT NULL, 
        [ORDER]           INT NOT NULL, 
        [ENABLED]          BIT
    );
    INSERT INTO #XMLItm
    (
        ID_OPTION_LIST, 
        ID_LIST, 
        CODE, 
        DESCRIPTION, 
        ORDER, 
        ENABLED
    )
           SELECT DISTINCT 
                  ROW_NUMBER() OVER (ORDER BY (SELECT NULL)) AS ID_OPTION_LIST,
                  -- I want delete this option and load the ID_LIST from #XMLLst
                  Ssn.value('(/item/@id_list)[1]', 'int') AS ID_LIST, 
                  Ssn.value('(/item/@code)[1]', 'varchar(50)') AS CODE, 
                  Ssn.value('(/item/@description)[1]', 'varchar(1000)') AS DESCRIPTION, 
                  Ssn.value('(/item/@order)[1]', 'int') AS ORDER, 
                  Ssn.value('(/item/@enabled)[1]', 'int') AS ENABLED
           FROM
           (
               SELECT S.s.query('.') AS Ssn
               FROM @xml.nodes('/lists/list/item') AS S(s)
           ) AS SSnes;
    WITH CTE_DATA_ORIGIN
         AS (SELECT DISTINCT 
                    ID_OPCION_LIST, 
                    ID_LIST, 
                    CODE, 
                    DESCRIPTION, 
                    ORDER, 
                    ENABLED
             FROM #XMLItm)
         MERGE INTO [sos].[ITEM_LIST] ITEM
         USING CTE_DATA_ORIGIN
         ON ITEM.ID_OPTION_LIST = CTE_DATA_ORIGIN.ID_OPTION_LIST
            AND ITEM.ID_LIST = CTE_DATA_ORIGIN.ID_LIST
             WHEN MATCHED AND CTE_DATA_ORIGIN.CODE = ITEM.CODE
             THEN UPDATE SET 
                             ITEM.DESCRIPTION = CTE_DATA_ORIGIN.DESCRIPTION, 
                             ITEM.ORDER = CTE_DATA_ORIGIN.ORDER, 
                             ITEM.ENABLED = CTE_DATA_ORIGIN.ENABLED
             WHEN NOT MATCHED
             THEN
               INSERT(ID_OPTION_LIST, 
                      ID_LIST, 
                      CODE, 
                      DESCRIPTION, 
                      ORDER, 
                      ENABLED)
               VALUES
         (
            CTE_DATOS_ORIGEN.ID_OPTION_LIST, 
            CTE_DATOS_ORIGEN.ID_LIST, 
            CTE_DATOS_ORIGEN.CODE, 
            CTE_DATOS_ORIGEN.DESCRIPTION, 
            CTE_DATOS_ORIGEN.ORDER, 
            CTE_DATOS_ORIGEN.ENABLED
         );
    DROP TABLE #XMLLst;
    DROP TABLE #XMLItm;

在这段代码中，我想在两个表中为一个公共字段 ID_LIST 加载信息。如何从临时表#XMLItm 中的#XMLLst 加载ID_LIST？

我想从表#XMLItm 中的XML 中删除字段ID_LIST，并从临时表#XMLLst 中获取ID_LIST；

功能：

在#XMLLst 表中加载一个包含您的 ID（自动递增）、名称和备注的列表，然后从 XML 导出到 SQL。

在表#XMLItm 中加载与id 列表链接的项目。

预期输出

+-----------+---------------+---------+
|ID_LIST    |    NAME_LIST  | REMARKS |
|1          |LIST A         | xx      |
|2          |LIST B         | yy      |
+-------------------------------------+

ID_OPTION_LIST  ID_LIST     CODE    DESCRIPTION ORDER   ENABLED
1               1           Y       yes         1       1
2               1           N       No          2       1
3               2           E       Enabled     1       1
4               2           D       Disables    2       1

【问题讨论】：

我真的不明白这篇文章的目的。 "...我想删除表中 ID_LIST 字段的冗余部分 #XMLItm..." ？！请编辑您的帖子并说明清楚。
Willmar，请通过 LinkedIn 或 Skype 与我联系。

标签： sql sql-server xml sql-server-2017 sql-server-2019

【解决方案1】：

请尝试以下解决方案。它将 list 与 XML 中的子 item 片段连接起来。此外，它还会在两个级别上生成两个运行序列。

SQL

-- DDL and sample data population, start
DECLARE @xml XML = 
'<?xml version="1.0" encoding="iso-8859-1"?>
<lists>
    <list list="LIST A" remark="xx">
        <item id_list="1" code="Y" description="Yes" order="1" enabled="1" />
        <item id_list="1" code="N" description="No" order="2" enabled="1" />
    </list>
    <list list="LIST B" remark="yy">
        <item id_list="2" code="E" description="Enabled" order="1" enabled="1" />
        <item id_list="2" code="D" description="Disabled" order="2" active="1" />
    </list>
</lists>';
-- DDL and sample data population, end

-- It seems exactly waht you need
SELECT ROW_NUMBER() OVER (ORDER BY i) AS ID_OPTION_LIST
    , DENSE_RANK() OVER (ORDER BY c) AS ID_LIST_Sequence
    , c.value('@list', 'VARCHAR(250)') AS NAME_LIST
    , c.value('@remark', 'VARCHAR(8000)') AS REMARKS
    , i.value('@id_list', 'int') AS ID_LIST
    , i.value('@code', 'varchar(50)') AS CODE 
    , i.value('@description', 'varchar(1000)') AS DESCRIPTION
    , i.value('@order', 'int') AS [ORDER]
    , i.value('@enabled', 'int') AS ENABLED
FROM @xml.nodes('/lists/list') AS t(c)
    CROSS APPLY t.c.nodes('item') AS x(i);

输出

+----------------+------------------+-----------+---------+---------+------+-------------+-------+---------+
| ID_OPTION_LIST | ID_LIST_Sequence | NAME_LIST | REMARKS | ID_LIST | CODE | DESCRIPTION | ORDER | ENABLED |
+----------------+------------------+-----------+---------+---------+------+-------------+-------+---------+
|              1 |                1 | LIST A    | xx      |       1 | Y    | Yes         |     1 | 1       |
|              2 |                1 | LIST A    | xx      |       1 | N    | No          |     2 | 1       |
|              3 |                2 | LIST B    | yy      |       2 | E    | Enabled     |     1 | 1       |
|              4 |                2 | LIST B    | yy      |       2 | D    | Disabled    |     2 | NULL    |
+----------------+------------------+-----------+---------+---------+------+-------------+-------+---------+

【讨论】：