SQL Server - XQuery - 子串

Question

What is the best way to retrieve values A and B on SQL Server? I have the following as a part of an XML type column.

<Data Name=""Teams"">
         <Data DataElement=""Test"" Value=""FirstTeam=A;SecondTeam=B;""/>
</Data>

Is it better to substring the query result (ie use sql substring) or is it more efficient to substring the xquery result (ie use substring in the xquery)?

Thank you

Answer 1

You can use normal xml.nodes to get that value from xml and then parse the value based on the requirement

declare @xml xml =
'<Data Name="Teams">
         <Data DataElement="Test" Value="FirstTeam=A;SecondTeam=B;"/>
</Data>'

select x.r.value(N'@Value', N'varchar(100)') as val from @xml.nodes('Data/Data') as x(r)

Answer 2

What does as a part of an XML type column mean?

• Your example is not valid XML (doubled " signs?)
• Your XML has a strange design (nested Data within Data ...)

Is there a reason not to put this like (much better to query!)?

<Teams>
  <Team DataElement="Test" Name="A"/>
  <Team DataElement="Test" Name="B"/>
</Teams>

As XML has - other than SQL Server - an inherent order by position, you would not need a "first" or "second" team. Well, of courrse you can add an attribute to specify this...

If this is a snippet of a big XML , you have to consider, that

• real XML will navigate down the tree only searching for < which is pretty fast
• multiple occurence of your target (would this be Value=" ?) will lead to unexpected results...
• Which data type do you use for storage? Hopefully this is XML ... For SUBSTRING you´d have to do an expensive cast...

So my clear advise: Use real XML methods to navigate to the given content and separate the content of @Value extra.

Try it like this to get the content:

DECLARE @XML XML=
'<Data Name="Teams">
         <Data DataElement="Test" Value="FirstTeam=A;SecondTeam=B;"/>
</Data>';
SELECT @XML.value('(//Data[@DataElement="Test"]/@Value)[1]','nvarchar(max)');

UPDATE Use a table and split it on-the-fly

DECLARE @tbl TABLE(YourXMLColumn XML);
INSERT INTO @tbl VALUES
(
'<Data Name="Teams">
         <Data DataElement="Test" Value="FirstTeam=A;SecondTeam=B;"/>
</Data>'
);

SELECT Team.value('.','nvarchar(max)') AS Team
FROM @tbl AS tbl
CROSS APPLY (SELECT CAST('<x>' + REPLACE(YourXMLColumn.value('(//Data[@DataElement="Test"]/@Value)[1]','nvarchar(max)'),';','</x><x>') + '</x>' AS XML)) AS Splitted(X)
CROSS APPLY X.nodes('x') AS The(Team)

The result

Team
-----------
FirstTeam=A
SecondTeam=B

(You might split this in two column again, but this is something you should solve on your own now :-).

The fact that you have to split one value to get entries you'd have to split again, shows clearly, that this is not the best design...