将每个子标签转换为具有多个分隔符的单个列-SQL Server(2)

时间:2017-06-11 06:37:45

标签: sql-server xml tsql xpath xquery

declare @xml xml='<plan>
    <prescriptions id="1">
        <prescription>
            <name>ABC</name>
            <frequency>Daily</frequency>
            <dailyfrequency>
               <morning>2</morning>
               <afternoon></afternoon>
               <night>1</night>
            </dailyfrequency>
            <dayfrequency></dayfrequency>
        </prescription>
        <prescription>
            <name>EDF</name>
            <frequency>Daily</frequency>
            <dailyfrequency>
              <morning>5</morning>
              <afternoon>5</afternoon>
              <night>1</night>
           </dailyfrequency>
           <dayfrequency></dayfrequency>
        </prescription>
        <prescription>
            <name>YTER</name>
            <frequency>Weekly</frequency>
            <dailyfrequency>
              <morning>5</morning>
              <afternoon>5</afternoon>
              <night>1</night>
            </dailyfrequency>
            <dayfrequency>Monday,Tuesday,Wednesday</dayfrequency>
        </prescription>
    </prescriptions>
    <prescriptions id="2">
        <prescription>
            <name>YTRE</name>
            <frequency>Daily</frequency>
            <dailyfrequency>
               <morning>2</morning>
               <afternoon></afternoon>
               <night>1</night>
            </dailyfrequency>
            <dayfrequency></dayfrequency>
          </prescription>
    </prescriptions>
</plan>'

我们可以像下面这样查询,用多个分隔符隔离每个元素,以产生相同的列。

SELECT STUFF(
(
    SELECT '!' + STUFF(p.query(N'for $n in .//*[local-name()!="dailyfrequency"]
                           return <a>{concat("$",($n/text())[1])}</a>'
                        ).value(N'.',N'nvarchar(max)'),1,1,'')
    FROM @xml.nodes(N'/plan/prescriptions/prescription') AS A(p)
    FOR XML PATH(''),TYPE).value(N'.',N'nvarchar(max)'),1,1,'')

结果:

ABC$Daily$2$$1$!EDF$Daily$5$5$1$!YTER$Weekly$5$5$1$Monday,Tuesday,Wednesday!YTRE$Daily$2$$1$

但问题是这将把所有标签组合在一个列下。当我们看xml时它有2个主要的分区处方id = 1和prescritpion id = 2.所以我们的最终结果将是这样的

结果:

ABC$Daily$2$$1$!EDF$Daily$5$5$1$!YTER$Weekly$5$5$1$Monday,Tuesday,Wednesday
YTRE$Daily$2$$1$

我认为我们必须为每个&lt; 处方&gt;切碎Xml然后必须计算这个

有人可以在这里解决这个问题

提前致谢,Jayendran

1 个答案:

答案 0 :(得分:1)

更改您的查询:

SELECT STUFF(
(
    SELECT '!' + STUFF(p.query(N'for $n in .//*[local-name()!="dailyfrequency"]
                           return <a>{concat("$",($n/text())[1])}</a>'
                        ).value(N'.',N'nvarchar(max)'),1,1,'')
    FROM p.nodes(N'prescription') AS A(p)
    FOR XML PATH(''),TYPE).value(N'.',N'nvarchar(max)'),1,1,'')
FROM @xml.nodes(N'/plan/prescriptions') AS A(p);

首先,我们为不同的处方创建一个派生表,其次我们分别使用每个处方的前代码。