Querying XML file in SQL Server - sql-server

Question

This is my xml file
<Detials xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">
<Parents>
<Parent id="1234">
<name>
<firstname>ABC</firstname>
<lastname>XYSX</lastname>
</name>
</Parent>
<Parent id="1235">
<name>
<firstname>TFU</firstname>
<lastname>GHY</lastname>
</name>
</Parent>
</Parents>
<Children>
<Child id="457" Parentid="1234">
<name>
<cfirstname>JOHN</cfirstname>
<clastname>SMITH</clastname>
</name>
</Child>
<Child id="459" Parentid="1235">
<name>
<cfirstname>DAVID</cfirstname>
<clastname>SMITH</clastname>
</name>
</Child>
</Children>
</Detials>
I stored it in a table (using Bulk Insert).
When I query like this
SELECT
x.c.value('(./Parents/Parent/#id)[1]', 'nvarchar(4)' ) AS id
FROM
T1 s
CROSS APPLY
s.XMLData.nodes('Detials') AS x(c)
I get the result as
Id
----
1234
When I changed it a little bit
SELECT
x.c.value('#id', 'nvarchar(4)' ) AS id
FROM
T1 s
CROSS APPLY
s.XMLData.nodes('/Detials/Parents/Parent') AS x(c)
I get:
Id
----
1234
1235
I just want to query them to result like this
Parent_id | firstname | lastname | Child_id | Child_First_name | child_last_name
----------+-----------+----------+----------+------------------+-----------
1234 | ABC | XYSX | 457 | JOHN | SMITH
1235 | TFU | GHY | 459 | DAVID | SMITH
How could I do this in query ?
In my result I want total no. of rows which depends on the Parentid. If my xml file contains 6 <parent id> then my query should show all 6 rows for each parent id's along with comparing the Childid's to also be included if the Parentid matches in <child> tag.
Thanks, Jayendran

Answer 1

You were on the right track, but this would require a left join on the two result sets Parent/Children
Example
Declare #T1 table (ID int,XMLData xml)
Insert Into #T1 values
(1,'<Detials xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"><Parents><Parent id="1234"><name><firstname>ABC</firstname><lastname>XYSX</lastname></name></Parent><Parent id="1235"><name><firstname>TFU</firstname><lastname>GHY</lastname></name></Parent></Parents><Children><Child id="457" Parentid="1234"><name><cfirstname>JOHN</cfirstname><clastname>SMITH</clastname></name></Child><Child id="459" Parentid="1235"><name><cfirstname>DAVID</cfirstname><clastname>SMITH</clastname></name></Child></Children></Detials>')
Select A.ID
,B.*
From #T1 A
Cross Apply (
Select B1.Parent_id
,B1.First_Name
,B1.Last_Name
,B2.Child_id
,B2.Child_First_name
,B2.child_last_name
From (
Select
Parent_id = x.c.value('#id', 'nvarchar(4)' )
,First_Name = x.c.value('(name/firstname)[1]','varchar(50)')
,Last_Name = x.c.value('(name/lastname)[1]','varchar(50)')
From #T1 s
Cross Apply s.XMLData.nodes('/Detials/Parents/Parent') AS x(c)
) B1
Left Join (
Select
pt_id = x.c.value('#Parentid', 'nvarchar(4)' )
,Child_id = x.c.value('#id', 'nvarchar(4)' )
,Child_First_name = x.c.value('(name/cfirstname)[1]','varchar(50)')
,child_last_name = x.c.value('(name/clastname)[1]','varchar(50)')
From #T1 s
Cross Apply s.XMLData.nodes('/Detials/Children/Child ') AS x(c)
) B2
on B1.Parent_id = B2.pt_id
) B
Returns

Answer 2

I'd solve this in one single go:
SELECT Parent.ID
,p.value('(name/firstname/text())[1]','nvarchar(max)')
,p.value('(name/lastname/text())[1]','nvarchar(max)')
,c.value('#id','int')
,c.value('(name/cfirstname/text())[1]','nvarchar(max)')
,c.value('(name/clastname/text())[1]','nvarchar(max)')
FROM #xml.nodes('/Detials/Parents/Parent') AS A(p)
OUTER APPLY (SELECT p.value('#id','int')) AS Parent(ID)
OUTER APPLY #xml.nodes('/Detials/Children/Child[#Parentid=sql:column("Parent.ID")]') AS B(c);
The trick is, to read the Parentid into a named column via APPLY. This value can be used as predicate to fetch the children via sql:column().