XML to SQL Server 2008 extract - sql-server-2008

Question

I need further assistance with extracting records from XML and loading it into a SQL Server table.
I have this as my #xml:
<admin submitter_id="login0" effective_date="mm/dd/yyyy">
<rec effected_id="login1" adjustment="100.00" type="foo">
<reason reason_id="1" />
<reason reason_id="2" />
</rec>
<rec effected_id="login2" adjustment="50.00" type="bar">
<reason reason_id="3" />
</rec>
</admin>
I need this from a result set:
login0, login1, mm/dd/yyyy, 100.00, foo, 1
login0, login1, mm/dd/yyyy, 100.00, foo, 2
login0, login2, mm/dd/yyyy, 50.00, bar, 3
Does that make sense? The adjustment to the reason_id is one to many. I have figured out how to extract all the values except for the second line. I can only obtain the first reason_id and then it proceeds to the next record. I think this can be beaten with a CROSS APPLY but I cannot get it to work. Please help!
oh, I may also have received bogus XML. So if that's wrong, please tell me!

Answer 1

How about something like
DECLARE #Xml XML
SELECT #Xml = '<admin submitter_id="login0" effective_date="mm/dd/yyyy">
<rec effected_id="login1" adjustment="100.00" type="foo">
<reason reason_id="1" />
<reason reason_id="2" />
</rec>
<rec effected_id="login2" adjustment="50.00" type="bar">
<reason reason_id="3" />
</rec>
</admin>'
SELECT #Xml,
A2.B.value('(../../#submitter_id)[1]','VARCHAR(50)'),
A2.B.value('(../#effected_id)[1]','VARCHAR(50)'),
A2.B.value('(../../#effective_date)[1]','VARCHAR(50)'),
A2.B.value('(../#adjustment)[1]','FLOAT'),
A2.B.value('(../#type)[1]','VARCHAR(50)'),
A2.B.value('(#reason_id)[1]','INT')
FROM #XML.nodes('//admin/rec/reason') A2(B)
SQL Fiddle DEMO

Answer 2

Try to use this T-SQL code snippet:
SELECT
Submitter = #xml.value('(/admin/#submitter_id)[1]', 'varchar(50)'),
EffectedID = Rec.value('(#effected_id)[1]', 'varchar(50)'),
DateStamp = #xml.value('(/admin/#effective_date)[1]', 'varchar(50)'),
TypeID = Rec.value('(#type)[1]', 'varchar(50)'),
ReasonID = Reason.value('(#reason_id)[1]', 'int')
FROM
#xml.nodes('/admin/rec') AS Tbl(Rec)
CROSS APPLY
Rec.nodes('reason') AS T2(Reason)
It gives me an output of:
You need to have two nested lists of nodes - the first one grabs all the <rec> nodes from inside <admin>, while the second one iterates over all the <reason> nodes inside each of the <rec> nodes. That way, you can reliably extract all information from those two nested levels of subnodes, no matter how many subnodes there are on each level.