extracting distinct attributes from a document

[Iain]

I've an xml document that looks a bit like this

<Vendors>
<Vendor Stationery="Fre d" />
<Vendor Stationery="ber t" />
<Vendor Stationery="ber t" />
</Vendors>

I want to extract a list of the distinct values of Stationery. I've found
some things on the web that suggest XPath statements to do this (in XSLT),
but I get the error 'The Expression passed to this should result in a node
set with my attempts. Here is my code

String XPathString = "Vendor/[not(@InvoiceSta tionery =
preceding-sibling::Vendor/@InvoiceStation ery]@InvoiceStation ery";

XmlNodeList Stationery = xmlVendors.Sele ctNodes(XPathSt ring);

I can select fine with Vendor/@InvoiceStation ery, but with the duplications.
I've tried several variations but I am unsure if it is the XPath or
something else. My GUESS is that the error means that the return is a
scalar value and not nodes, but I don't find it the clearest description
ever!

Iain

[Oleg Tkachenko [MVP]]

Iain wrote:

<Vendors>
<Vendor Stationery="Fre d" />
<Vendor Stationery="ber t" />
<Vendor Stationery="ber t" />
</Vendors>

I want to extract a list of the distinct values of Stationery. I've found
some things on the web that suggest XPath statements to do this (in XSLT),
but I get the error 'The Expression passed to this should result in a node
set with my attempts. Here is my code

String XPathString = "Vendor/[not(@InvoiceSta tionery =
preceding-sibling::Vendor/@InvoiceStation ery]@InvoiceStation ery";

Well,
1) in your XML there is no InvoiceStatione ry attribute.
2) several syntax errors, try
String XPathString = "Vendor/[not(@Stationery =
preceding-sibling::Vendor/@Stationery)]/@Stationery";

Beware, that's goind to be deadly slow on big XML documents.
Much scalable solution is to leverage EXSLT.NET [1] set:distinct
extension function.
[1] http://workspaces.gotdotnet.com/exslt
--
Oleg Tkachenko [XML MVP, XmlInsider]
http://blog.tkachenko.com

[Iain]

Thanks, Oleg. In real life I had the attribute InvoiceStatione ry, but
shortened it for the example. For simplicity I've now set the attribute
name to be Stationery and provided a 3 line sample ...

But. Your code copied as is does not work. It does the same thing.

I'm doing this with an XmlDocument - that should be OK, should it not?

XmlDocument xx = new XmlDocument();
xx.LoadXml("<Ve ndors><Vendor Stationery=\"Fr ed\" /> <Vendor
Stationery=\"be rt\" /> <Vendor Stationery=\"be rt\" /></Vendors>");

String XPathString = "Vendor/[not(@Stationery =
preceding-sibling::Vendor/@Stationery)]/@Stationery";

XmlNodeList test = xx.SelectNodes( XPathString);

Thanks

iain
"Oleg Tkachenko [MVP]" <oleg@NO!SPAM!P LEASEtkachenko. com> wrote in message
news:#X******** ******@TK2MSFTN GP09.phx.gbl...

Well,
1) in your XML there is no InvoiceStatione ry attribute.
2) several syntax errors, try
String XPathString = "Vendor/[not(@Stationery =
preceding-sibling::Vendor/@Stationery)]/@Stationery";

Beware, that's goind to be deadly slow on big XML documents.
Much scalable solution is to leverage EXSLT.NET [1] set:distinct
extension function.
[1] http://workspaces.gotdotnet.com/exslt
--
Oleg Tkachenko [XML MVP, XmlInsider]
http://blog.tkachenko.com

[Oleg Tkachenko [MVP]]

Iain wrote:

Thanks, Oleg. In real life I had the attribute InvoiceStatione ry, but
shortened it for the example. For simplicity I've now set the attribute
name to be Stationery and provided a 3 line sample ...

But. Your code copied as is does not work. It does the same thing.

I'm doing this with an XmlDocument - that should be OK, should it not?

XmlDocument xx = new XmlDocument();
xx.LoadXml("<Ve ndors><Vendor Stationery=\"Fr ed\" /> <Vendor
Stationery=\"be rt\" /> <Vendor Stationery=\"be rt\" /></Vendors>");

String XPathString = "Vendor/[not(@Stationery =
preceding-sibling::Vendor/@Stationery)]/@Stationery";

Syntax error again. It should be
Vendor[not(@Stationery =
preceding-sibling::Vendor/@Stationery)]/@Stationery"

And as this expression starts with Vendor, it must be executed on
Vendors element:
xx.DocumentElem ent.SelectNodes (XPathString);

Then it gives me 2 distinct Stationery attributes.
--
Oleg Tkachenko [XML MVP, XmlInsider]
http://blog.tkachenko.com

[Iain]

Thanks very much. I'm still struggling with XPath - not really getting my
mind round it.

In the meantime, I just selected all the nodes, copied the values to an
array, sorted it and skipped the duplicates (incidentally, I was amazed at
how easy this was to do. Try doing this in C++ in roughly 30 seconds!). I
feel ashamed, but it works and may even be more efficient!

But I'm glad to know how I should have done it!

Iain

"Oleg Tkachenko [MVP]" <oleg@NO!SPAM!P LEASEtkachenko. com> wrote in message
news:ui******** ******@TK2MSFTN GP12.phx.gbl...

Iain wrote:

Thanks, Oleg. In real life I had the attribute InvoiceStatione ry, but
shortened it for the example. For simplicity I've now set the attribute
name to be Stationery and provided a 3 line sample ...

But. Your code copied as is does not work. It does the same thing.

I'm doing this with an XmlDocument - that should be OK, should it not?

XmlDocument xx = new XmlDocument();
xx.LoadXml("<Ve ndors><Vendor Stationery=\"Fr ed\" /> <Vendor
Stationery=\"be rt\" /> <Vendor Stationery=\"be rt\" /></Vendors>");

String XPathString = "Vendor/[not(@Stationery =
preceding-sibling::Vendor/@Stationery)]/@Stationery";

Syntax error again. It should be
Vendor[not(@Stationery =
preceding-sibling::Vendor/@Stationery)]/@Stationery"

And as this expression starts with Vendor, it must be executed on
Vendors element:
xx.DocumentElem ent.SelectNodes (XPathString);

Then it gives me 2 distinct Stationery attributes.
--
Oleg Tkachenko [XML MVP, XmlInsider]
http://blog.tkachenko.com