Why does this xpath fail?

[David Thielen]

XPathNavigator nav = MyCreateNav(); // InnerXml == "software"
nav.SelectSingleNode".[.='software']");

The select returns an exception:
+ $exception {"'.[.='software']' has an invalid token."} System.Exception
{System.Xml.XPath.XPathException}

Any idea why?

--
thanks - dave
david_at_windward_dot_net
http://www.windwardreports.com

[Martin Honnen]

David Thielen wrote:

XPathNavigator nav = MyCreateNav(); // InnerXml == "software"
nav.SelectSingleNode".[.='software']");

^^^^^^
That is not even syntactically correct C# so first post the real code
then we can discuss the XPath expression.
--

Martin Honnen --- MVP XML
http://JavaScript.FAQTs.com/

[David Thielen]

Sorry - dropped a (

XPathNavigator nav = MyCreateNav(); // InnerXml == "software"
nav.SelectSingleNode(".[.='software']");

--
thanks - dave
david_at_windward_dot_net
http://www.windwardreports.com

"Martin Honnen" wrote:

David Thielen wrote:

XPathNavigator nav = MyCreateNav(); // InnerXml == "software"
nav.SelectSingleNode".[.='software']");

^^^^^^
That is not even syntactically correct C# so first post the real code
then we can discuss the XPath expression.
--

Martin Honnen --- MVP XML
http://JavaScript.FAQTs.com/

[Martin Honnen]

David Thielen wrote:

XPathNavigator nav = MyCreateNav(); // InnerXml == "software"
nav.SelectSingleNode(".[.='software']");

It is not a syntatically correct XPath expression, see
<http://www.w3.org/TR/xpath#section-Location-Steps>
which defines

Step ::= AxisSpecifier NodeTest Predicate*
| AbbreviatedStep

what you have with . is an abbreviated step, there is no predicate
allowed after that.
--

Martin Honnen --- MVP XML
http://JavaScript.FAQTs.com/

[David Thielen]

If I want to test if the node == "software", what is the syntax to do that?

--
thanks - dave
david_at_windward_dot_net
http://www.windwardreports.com

"Martin Honnen" wrote:

David Thielen wrote:

XPathNavigator nav = MyCreateNav(); // InnerXml == "software"
nav.SelectSingleNode(".[.='software']");

It is not a syntatically correct XPath expression, see
<http://www.w3.org/TR/xpath#section-Location-Steps>
which defines

Step ::= AxisSpecifier NodeTest Predicate*
| AbbreviatedStep

what you have with . is an abbreviated step, there is no predicate
allowed after that.
--

Martin Honnen --- MVP XML
http://JavaScript.FAQTs.com/

[Martin Honnen]

David Thielen wrote:

If I want to test if the node == "software", what is the syntax to do that?

self::node()[.='software']
would work with SelectSingleNode or Select as it finds a node set.
But of course
. = 'software'
is an allowed expression returning a boolean value, so with the XPath
API you would need the Evaluate method and not SelectSingleNode/Select.
--

Martin Honnen --- MVP XML
http://JavaScript.FAQTs.com/

[Cerebrus99]

Hi,

In other words, David, first of all, you must decide if you want to *select*
the node or do you just want to *check* if that particular node has this
value. The implementation would differ in both cases.

Assuming you want to select only 1 node, use an
XmlDocument.SelectSingleNode("XpathExpression"). You can also use an XmlNode
in place of a XmlDocument. (To correct your code, the XPathNavigator class
does not have any method called "SelectSingleNode")

If you want to select multiple nodes and must use a XPathNavigator, then use
nav.Select("XpathExpression").

If you want to check for a value, use
myReturnValue = CType(nav.Evaluate("XpathExpression"), Type)

As far as the XPath expression to select a particular node is concerned,
that depends on your Xml. However, I have read that the more exactly you
define your expression, the better the performance.

For instance,
/ROOTNODE/CHILDNODE/GRANDCHILDNODE[.='software']

If you share your Xml format, it might be easier to define the exact Xpath
to use.

Hope this helps,

Regards,

Cerebrus.
"Martin Honnen" <ma*******@yahoo.de> wrote in message
news:uN**************@TK2MSFTNGP09.phx.gbl...

David Thielen wrote:

If I want to test if the node == "software", what is the syntax to do

that?
self::node()[.='software']
would work with SelectSingleNode or Select as it finds a node set.
But of course
. = 'software'
is an allowed expression returning a boolean value, so with the XPath
API you would need the Evaluate method and not SelectSingleNode/Select.
--

Martin Honnen --- MVP XML
http://JavaScript.FAQTs.com/

[Martin Honnen]

Cerebrus99 wrote:

(To correct your code, the XPathNavigator class
does not have any method called "SelectSingleNode")

It has in .NET 2.0.

--

Martin Honnen --- MVP XML
http://JavaScript.FAQTs.com/

[David Thielen]

Hi;

Here is what I am doing. The xml is:
<order>
<product>software</product>
</order>

If I create an
XPathNavigator nav = XPathNavigator("/order"); // actually created using
XPathDocument and selecting down to that node
I can call:
nav.SelectSingleNode("product[.='software']");

But these fail (the selects):
XPathNavigator nav = XPathNavigator("/order/product");
nav.SelectSingleNode("[.='software']");
nav.SelectSingleNode(".[.='software']");
nav.SelectSingleNode("./[.='software']");

Why? It seems to me this should return the node <product>software</product>.
The only thing that works for this nav is:
nav.Evaluate(".='software'");

I am probably not understanding something here but I have no idea what and I
have spent a couple of hours trying to figure this out. Help please.

--
thanks - dave
david_at_windward_dot_net
http://www.windwardreports.com

[Cerebrus]

Hi David,

Wow ! You took almost a week to reply back !

I'm not using VS 2005, so not familiar with the SelectSingleNode method of
the XPathNavigator class, so I will assume, that all it requires is a string
containing the XPathExpression, and in that case, the problem is not with
the method, but it is with the XPathExpression you're using.

By using square brackets in an XPath expression you specify an element
further. So you cannot simply use a condition in square brackets, and expect
something to be selected. It has to be in the form :

/Node[condition]

(Select a node where this condition holds true)

While nav.Evaluate(".='software'") checks if the current node is "software".
Since the current node is at the start of the document by default, it
returns a Boolean False value. It works, but returns False, and doesn't do
what you intend to do.

So, my final suggestion is just use the following Xpath expression (looking
at your 3 line Xml sample)

nav.SelectSingleNode("order/product[.='software']")

Regards,

Cerebrus.


"David Thielen" <da***@bogus.windward.net> wrote in message
news:F0**********************************@microsof t.com...

Hi;

Here is what I am doing. The xml is:
<order>
<product>software</product>
</order>

If I create an
XPathNavigator nav = XPathNavigator("/order"); // actually created using
XPathDocument and selecting down to that node
I can call:
nav.SelectSingleNode("product[.='software']");

But these fail (the selects):
XPathNavigator nav = XPathNavigator("/order/product");
nav.SelectSingleNode("[.='software']");
nav.SelectSingleNode(".[.='software']");
nav.SelectSingleNode("./[.='software']");

Why? It seems to me this should return the node <product>software</product>. The only thing that works for this nav is:
nav.Evaluate(".='software'");

I am probably not understanding something here but I have no idea what and I have spent a couple of hours trying to figure this out. Help please.

--
thanks - dave
david_at_windward_dot_net
http://www.windwardreports.com

[David Thielen]

Hello;

Unfortunately I cannot do that. This is for a product where it is passed an
xpath to get a list of nodes ("/order/product") and then when iterating
through that list I have to see if nodes exist for an iteration of the list
(.='software').

And the inner check to see if a node exists could be a node, a list of
nodes, nodes with children, or just ".='software'". This works fine using
dom4j/jaxen in the java world - it returns a single node in this case.

And it seems to me there should be a way to do this...

--
thanks - dave
david_at_windward_dot_net
http://www.windwardreports.com

"Cerebrus" wrote:

Hi David,

Wow ! You took almost a week to reply back !

I'm not using VS 2005, so not familiar with the SelectSingleNode method of
the XPathNavigator class, so I will assume, that all it requires is a string
containing the XPathExpression, and in that case, the problem is not with
the method, but it is with the XPathExpression you're using.

By using square brackets in an XPath expression you specify an element
further. So you cannot simply use a condition in square brackets, and expect
something to be selected. It has to be in the form :

/Node[condition]

(Select a node where this condition holds true)

While nav.Evaluate(".='software'") checks if the current node is "software".
Since the current node is at the start of the document by default, it
returns a Boolean False value. It works, but returns False, and doesn't do
what you intend to do.

So, my final suggestion is just use the following Xpath expression (looking
at your 3 line Xml sample)

nav.SelectSingleNode("order/product[.='software']")

Regards,

Cerebrus.


"David Thielen" <da***@bogus.windward.net> wrote in message
news:F0**********************************@microsof t.com...

Hi;

Here is what I am doing. The xml is:
<order>
<product>software</product>
</order>

If I create an
XPathNavigator nav = XPathNavigator("/order"); // actually created using
XPathDocument and selecting down to that node
I can call:
nav.SelectSingleNode("product[.='software']");

But these fail (the selects):
XPathNavigator nav = XPathNavigator("/order/product");
nav.SelectSingleNode("[.='software']");
nav.SelectSingleNode(".[.='software']");
nav.SelectSingleNode("./[.='software']");

Why? It seems to me this should return the node

<product>software</product>.

The only thing that works for this nav is:
nav.Evaluate(".='software'");

I am probably not understanding something here but I have no idea what and

I

have spent a couple of hours trying to figure this out. Help please.

--
thanks - dave
david_at_windward_dot_net
http://www.windwardreports.com

[Cerebrus]

Hi David,

Unfortunately I cannot do that. This is for a product where it is passed an
xpath to get a list of nodes ("/order/product") and then when iterating
through that list I have to see if nodes exist for an iteration of the list
(.='software')
I don't see the problem. If you mean, you cannot fully define the
XPathExpression,
you did mention that nav.SelectSingleNode("product[.='software']");
works for you.
Use Wildcards if necessary.

//product[.='software'] will select all Product Nodes which match the
condition (even if at different levels)
And the inner check to see if a node exists could be a node, a list of
nodes, nodes with children, or just ".='software'". This works fine using
dom4j/jaxen in the java world - it returns a single node in this case.

Use checks using the XmlNodeType enumeration for each of the
possibilities.

HTH,

Regards,

Cerebrus.

[David Thielen]

Hi;

The problem we have is our customers enter the xpath in pieces and we have
no control over it. I know what is happening in this specific case because I
have a set of xpath a customer uses that breaks here.

But I can't change what they enter and I can't control it. In this case the
problem I am facing is "product[.='software']" works but ".[.='software']" is
illegal and I am trying to figure out how to handle this.

And the root problem I have is I don't understand why ".[.='software']" is
illegal - it seems to me that it should return that one node.

--
thanks - dave
david_at_windward_dot_net
http://www.windwardreports.com

"Cerebrus" wrote:

Hi David,

Unfortunately I cannot do that. This is for a product where it is passed an
xpath to get a list of nodes ("/order/product") and then when iterating
through that list I have to see if nodes exist for an iteration of the list
(.='software')

I don't see the problem. If you mean, you cannot fully define the
XPathExpression,
you did mention that nav.SelectSingleNode("product[.='software']");
works for you.
Use Wildcards if necessary.

//product[.='software'] will select all Product Nodes which match the
condition (even if at different levels)

And the inner check to see if a node exists could be a node, a list of
nodes, nodes with children, or just ".='software'". This works fine using
dom4j/jaxen in the java world - it returns a single node in this case.

Use checks using the XmlNodeType enumeration for each of the
possibilities.

HTH,

Regards,

Cerebrus.

[David Thielen]

Hi;

I reposted this as a new question with a sample program to give a real clear
example.

--
thanks - dave
david_at_windward_dot_net
http://www.windwardreports.com