如何在 SQL Server 上使用 XQuery 修改 xml 子节点

Question

For an xml field (SQL Server) I need to add a node in every subnode, based on a source table and a condition. This is my xml data:

declare @X table (XMLDATA xml)
insert @X values('
<row>
  <node>
    <name>Francesco</name>
  </name>
  <node>
    <label>Alessandro</name>
  </node>
  <node>
    <name>Daniele</name>
  </node>
</row>')

For every <name> , I want to add a node <number> . The matches for numbers and names are written in a table @T :

declare @T table (name varchar(20), number int)
insert @T values
('Alessandro', 24)
,('Francesco', 10)
,('Daniele', 16)

To update the nodes I use XMLDATA.modify , and I use xpath conditions to select the right node:

update @X set XMLDATA.modify('insert element number {sql:column("number")} as last into (row/node[name=sql:column("name")])[1]')
from @X
cross join @T

The query above works only for the first row of @T (in the example is Alessandro/24). The other 2 rows of @T are ignored. I need to add number to every node . This is the final XMLDATA:

<row>
  <node>
    <name>Francesco</name>
  </node>
  <node>
    <name>Alessandro</name>
    <number>24</number>
  </node>
  <node>
    <name>Daniele</name>
  </node>
</row>

Answer 1

I use while loop. Please check below code, it may be help you.

declare @X table (XMLDATA xml)
insert @X values('
<row>
  <node>
    <name>Alessandro</name>
  </node>
  <node>
    <name>Francesco</name>
  </node>
  <node>
    <name>Daniele</name>
  </node>
</row>')


declare @T table (name1 varchar(20), number int,RowID int identity(1,1) not null)
insert @T values
('Alessandro', 24)
,('Francesco', 10)
,('Daniele', 16)

DECLARE @i int,@iCount int,@namevalue varchar(100)='',@number int
SET @i = 1


SELECT @iCount=MAX(RowID) FROM @T

WHILE (@i <= @iCount)
BEGIN

SELECT @namevalue=name1 FROM @T WHERE RowID=@i
SELECT @number=number FROM @T WHERE RowID=@i

update @X set XMLDATA.modify('insert element number {sql:variable("@number")} as last into (row/node)[name=sql:variable("@namevalue")][1]')
from @X
cross join @T

  SET @i = @i + 1
END

SELECT * FROM @X

Output :

<row>
  <node>
    <name>Alessandro</name>
    <number>24</number>
  </node>
  <node>
    <name>Francesco</name>
    <number>10</number>
  </node>
  <node>
    <name>Daniele</name>
    <number>16</number>
  </node>
</row>

Thanks .

Answer 2

If your real life data is as simple as the example, the solution can be found much easier:

declare @X table (XMLDATA xml)
insert @X values(
'<row>
  <node>
    <name>Francesco</name>
  </node>
  <node>
    <name>Alessandro</name>
  </node>
  <node>
    <name>Daniele</name>
  </node>
</row>')

declare @T table (name varchar(20), number int)
insert @T values
('Alessandro', 24)
,('Francesco', 10)
,('Daniele', 16);

SELECT nm.value(N'.','nvarchar(max)') AS name
        ,t.number 
FROM @X
CROSS APPLY XMLDATA.nodes(N'/row/node/name') AS A(nm)
INNER JOIN @T AS t ON t.name=nm.value(N'.','nvarchar(max)')
FOR XML PATH('node'),ROOT('row')

If this does not work for you, please provide more realistic sample data!