How to retrieve data from XML column in SQL - sql

Question

I have the following table layout:
OrderDetails:
ItemID (PK, int, not null)
ItemName (nvarchar(450), null)
OrderID (FK, int, not null)
Discounts (nvarchar(max), null)
The Discounts column is declared nvarchar(max) but the data is really XML - no idea why it wasn't declared as XML. I need to query the table to show the OrderID, ItemName, prorated discounts, all other discounts, and total of prorated + all other discounts. Here is an example of some records.
ItemID ItemName OrderID Discounts
8610 Item 1 4227 SEE XML 4227 BELOW
8615 Item 2 4227 <DocumentElement></DocumentElement> //no discounts for this row
8620 Item 3 9387 SEE XML 9387 BELOW
XML OrderId = 4227:
<DocumentElement>
<DiscountsTable>
<DiscountDisplayName>Bundle A</DiscountDisplayName>
<DiscountValue>6.00</DiscountValue>
</DiscountsTable>
<DiscountsTable>
<DiscountDisplayName>Bundle B</DiscountDisplayName>
<DiscountValue>25.00</DiscountValue>
</DiscountsTable>
</DocumentElement>
XML for OrderId = 9387:
<DocumentElement>
<DiscountsTable>
<DiscountDisplayName>Prorated Discount</DiscountDisplayName>
<DiscountValue>6.45</DiscountValue>
</DiscountsTable>
<DiscountsTable>
<DiscountDisplayName>Bundle A</DiscountDisplayName>
<DiscountValue>5.61</DiscountValue>
</DiscountsTable>
<DiscountsTable>
<DiscountDisplayName>Bundle B</DiscountDisplayName>
<DiscountValue>23.39</DiscountValue>
</DiscountsTable>
</DocumentElement>
So, What I need is a query that will return the ItemID, ItemName, aggregated prorated discounts, aggregated bundled discounts, and a total of the discounts added together (prorated + bundled = total discounts). For the 3 records above, the query result should look like this:
Item ID Item Name Prorated Discounts Other Discounts Total Discounts
8610 Item 1 0.00 31.00 31.00
8615 Item 2 0.00 0.00 0.00
8620 Item 3 6.45 29.00 35.45
I've tried using the value() method but I get an error stating it can not be used with nvarchar(max). How do I parse the XML column to pull aggregated values?

Answer 1

You need to convert the column to XML first to use SQLXML features. Something like this.
;with tbl as (
select ItemID, ItemName, OrderID,
convert(xml, Discounts) as Discounts
from OrderDetails
)
select ItemID, ItemName, OrderID,
t.v.value('DiscountDisplayName[1]','varchar(100)') DiscountDisplayName,
t.v.value('DiscountValue[1]','float') DiscountValue
from tbl cross apply Discounts.nodes('DocumentElement/DiscountsTable') t(v)
Elaborate a little
;with tbl as (
select ItemID, ItemName, OrderID,
convert(xml, Discounts) as Discounts
from OrderDetails
),
tbl1 as (
select ItemID, ItemName, OrderID,
t.v.value('(DiscountsTable[DiscountDisplayName="Prorated Discount"]/DiscountValue)[1]','float') Prorated, -- filter
t.v.value('(DiscountsTable[DiscountDisplayName="Bundle A"]/DiscountValue)[1]','float') BundleA,
t.v.value('(DiscountsTable[DiscountDisplayName="Bundle B"]/DiscountValue)[1]','float') BundleB
from tbl cross apply Discounts.nodes('DocumentElement') t(v) --do not go deeper
)
select ItemID, ItemName, OrderID, isnull(Prorated, 0) Prorated,
isnull(BundleA, 0) + isnull(BundleB, 0) Other,
isnull(Prorated, 0) + isnull(BundleA, 0) + isnull(BundleB, 0) Total
from tbl1

Answer 2

I would solve it as follows:
WITH Conv AS
(
SELECT ItemID,ItemName,OrderID,CONVERT(xml,Discounts) XmlVal
FROM Src
)
SELECT ItemID, ItemName, OrderID, Prorated, Other, Prorated+Other Total
FROM Conv
CROSS APPLY
(
SELECT SUM(CASE X.exist('DiscountDisplayName[text() = "Prorated Discount"]')
WHEN 1 THEN X.value('DiscountValue[1]', 'decimal(15,2)')
ELSE 0 END) Prorated,
SUM(CASE X.exist('DiscountDisplayName[text() = "Prorated Discount"]')
WHEN 0 THEN X.value('DiscountValue[1]', 'decimal(15,2)')
ELSE 0 END) Other
FROM XmlVal.nodes('/DocumentElement/DiscountsTable') T(X)
) T(Prorated, Other)
Results:
ItemID ItemName OrderID Prorated Other Total
----------- -------- ----------- ------------ -------- ---------
8610 Item 1 4227 0.00 31.00 31.00
8615 Item 2 4227 NULL NULL NULL
8620 Item 3 9387 6.45 29.00 35.45
You can add COALESCE or ISNULL to remove NULLs.