This is my last posting on my "Typed XML" blog. After almost 9 years at Microsoft I felt the need to do something else and on August 13th I joined the ranks of a well known online retailer. You'll still find some of my contributions in SQL Server 2008 when it is released. The entire team did a great job addressing some of the issues most often mentioned by customers. I don't want to steal anybody's thunder (or attract the attention of any lawyers :-) ) so I won't say much more but keep an eye out for that next version. I's going to rock!

Those 8 and 3/4 years at Microsoft were quite a ride. Thanks to everybody who helped make that time special. 

Here’s what the poster was doing

declare @xml xml

set @xml = '<root><test></test></root>'

set @xml.modify('replace value of (/root/test/text())[1] with "test

new value"')

select @xml

This user expected to get something like <root><test>test new value</test></root> but instead, the resulting instance looked like <root><test /></root>.

The server never gave any error or warning, yet the XML instance appeared to be untouched.

The problem, as you might have guessed already, is that XPath expression (/root/test/text())[1] returns the empty sequence. In the original instance, ‘test’ is an empty element. It has no children at all. Therefore the expression (/root/test/text())[1] doesn’t point to any existing node. The solution to this problem is to insert a new test node inside the ‘test’ element, like this

set @xml.modify('insert text{"test new value"} as first into (/root/test)[1] ')

In general, when you use the modify() method and the target XPath expression returns the empty sequence, nothing happens. The only exception to this rule is when the server can statically determine that the result will be empty. That’s when XML schemas come in handy.

For example, let’s look at the following example

DECLARE @x XML

SET @x = '<root><x/><y/><z/></root>'

SET @x.modify('delete /root[1]/a[1]')

Let’s suppose I made a typo and instead of typing ‘z’ in my XPath expression I typed ‘a’. The query doesn’t do anything because with this particular XML instance the XPath expression /root[1]/a[1] returns the empty sequence (it doesn’t point to any existing node). There is no way to know that before executing the query. Statically, the type of this XPath expression is element(a, xdt:untyped)?

Now, let’s try something similar when a schema collection is associated with the XML variable.

CREATE XML SCHEMA COLLECTION SC AS '

<schema xmlns="https://www.w3.org/2001/XMLSchema">

      <element name="root">

            <complexType>

                  <sequence>

                        <element name="x" type="string" minOccurs="0"/>

                        <element name="y" type="string" minOccurs="0"/>

                        <element name="z" type="string" minOccurs="0"/>

                  </sequence>

            </complexType>

      </element>

</schema>'

go

DECLARE @x XML(SC)

SET @x = '<root><x/><y/><z/></root>'

SET @x.modify(' delete /root[1]/a[1]')

This time the query fails with the following error message:

XQuery [modify()]: There is no element named 'a' in the type 'element(root,#anonymous) ?'.

The server was able to use the schema to determine that my XPath expression will never point to an existing node and sent me a helpful error message. Even though typing your data doesn’t mean you’ll catch all those mistakes, it still helps when debugging XML queries.

-
Disclaimer:
This posting is provided “AS IS” with no warranties, and confers no rights.
Use of included script samples are subject to the terms specified at https://www.microsoft.com/info/cpyright.htm.

This time we will look at DML operations on mixed content.  We’ll keep using the XML schema as well as the table I created in the first post of this series.

Let’s insert an instance to work with.

INSERT INTO T VALUES (3, CONVERT(XML(SC_Letter), '<l:letter xmlns:l="urn:letter" xmlns:t="urn:letter:tags">

      <l:date>2006-10-15-08:00</l:date>

      <l:sender>

            <l:FirstName>James</l:FirstName>

            <l:LastName>Kirk</l:LastName>

            <l:Line1>1234 Elm Street</l:Line1>

            <l:City>Redmond</l:City>

            <l:ZIPCode>98052</l:ZIPCode>

            <l:State>WA</l:State>

      </l:sender>

      <l:recipient>

            <l:FirstName>Jean-Luc</l:FirstName>

            <l:LastName>Picard</l:LastName>

            <l:Line1>13 rue des Lilas</l:Line1>

            <l:City>Ronchin</l:City>

            <l:ZIPCode>59555</l:ZIPCode>

            <l:Country>France</l:Country>

      </l:recipient>   

Dear Sir,

      I received your letter dated <t:date>2006-10-10+00:00</t:date> this morning. It would be my pleasure to attend the annual symposium <t:title>XML database systems in class-B Federation Spacecrafts</t:title>.

Please contact my assistant <t:person>Mr. Spock</t:person> at <t:email>spock@space-federation.org</t:email> to make all necessary arrangements.

Sincerely,

James T. Kirk

</l:letter>', 1))

go

DML operations on elements

There are no behavior differences between element-only content and mixed content. You can insert, delete, and replace the value of elements provided the resulting instance validates against the schema.

For example, let’s update the name of the assistant mentioned in the letter.

UPDATE T SET xmlCol.modify(' declare namespace l = "urn:letter";

declare namespace t = "urn:letter:tags";

replace value of (/l:letter//t:person[data(.) = "Mr. Spock"])[1] with "Capt. Spock"')

WHERE iCol = 3

go

You can query the instance back and see for yourself that the value of the element has been updated.

We could also decide not to mention the assistant by name and delete the element.

UPDATE T SET xmlCol.modify(' declare namespace l = "urn:letter";

declare namespace t = "urn:letter:tags";

delete (/l:letter//t:person[data(.) = "Capt. Spock"])[1]')

WHERE iCol = 3

go

If you query the instance back you can see that the letter now says

Please contact my assistant  at <t:email>spock@space-federation.org</t:email> to make all necessary arrangements.

Something very interesting happened here that is not immediately visible. If you remember, in our previous posts, we used the following query to examine the nodes in our instance

SELECT  x.query('.'), x.query('if (. instance of text()) then "text" else

      if (. instance of element()) then "element" else ""')

FROM T

CROSS APPLY xmlCol.nodes('/*:letter[1]/node()') AS Tref(x)

WHERE iCol = 3

Had we run this query prior to deleting the <person> element, we would’ve seen this as part of our results (partial results printed to save space)

Now that we’ve deleted the element, the results look like this (partial results again)

Since we delete an element that was between two text nodes, those nodes merged to form only one.

DML operations on text nodes

Text nodes can be updated, inserted or deleted.

First, let’s replace the value of a text node.

UPDATE T SET xmlCol.modify(' declare namespace l = "urn:letter";

declare namespace t = "urn:letter:tags";

replace value of /l:letter[1]/text()[6] with ". I suggest that you talk to my asisstant, who can be contacted at "')

WHERE iCol = 3

go

You can query the instance to see that the text of the letter was modified appropriately.

Now let’s insert a text node after an existing one

UPDATE T SET xmlCol.modify(' declare namespace l = "urn:letter";

declare namespace t = "urn:letter:tags";

insert text{"&#xD;&#x9;&#xD;&#x9;PS: Live long and prosper."} as last into /l:letter[1]')

WHERE iCol = 3

go

If you look at the nodes, you’ll see that the new text node we inserted merged with the one preceding it to form one text node so that all text between two elements tags counts as only one text node (partial results printed below)

I started working on this two days ago. I quickly realized that when I’m done I’ll most likely have material for multiple posts. My original plan was to finish the article in the next few days, and then publish it as several posts over time. However I eventually decided to start posting now, while the work is still in progress, and let any feedback or questions you might have guide me.

Description

Mixed content is one of the great but often overlooked features of xml schema. A mixed complex type makes it legal for text nodes to appear between the elements that are normally expected by the validator.

In order for mixed content to be allowed the <xsd:complexType> node of the type's declaration must contain an attribute named "mixed" with a value of "true".

For a more complete explanation you can start with the Section of the W3C's xsd primer that deals with mixed content at https://www.w3.org/TR/xmlschema-0/#mixedContent.

In this post we will examine SQL Server 2005's capabilities and behavior when dealing with mixed content.

Example

As an example, we’ll deal with letters. Letters are ideal candidates for mixed content because they can contain both structured and unstructured information. The sender's and recipient(s) addresses for example constitute structured data for which we can design content models. The body of a letter itself is free-flow text but within it we can identify certain valuable pieces of information such as dates or email addresses using XML elements.

Here is a sample schema for letters.

CREATE XML SCHEMA COLLECTION SC_Letter AS '

<xsd:schema xmlns:xsd="https://www.w3.org/2001/XMLSchema"

targetNamespace="urn:letter"

elementFormDefault="qualified"

xmlns="urn:letter">

      <xsd:element name="letter">

            <xsd:complexType mixed="true">

                  <xsd:sequence>

                        <xsd:element name="date" type="xsd:date"/>

                        <xsd:element name="sender" type="addressType"/>

                        <xsd:sequence minOccurs="1" maxOccurs="unbounded">

                              <xsd:element name="recipient" type="addressType"/>

                        </xsd:sequence>

                        <xsd:any processContents="strict" namespace="urn:letter:tags" minOccurs="0" maxOccurs="unbounded"/>

                  </xsd:sequence>

            </xsd:complexType>

      </xsd:element>

      <xsd:complexType name="addressType">

            <xsd:sequence>

                        <xsd:element name="Title" type="xsd:string" minOccurs="0"/>

                        <xsd:element name="FirstName" type="xsd:string" minOccurs="0"/>

                        <xsd:element name="LastName" type="xsd:string"/>

                        <xsd:element name="Line1" type="xsd:string"/>

                        <xsd:element name="Line2" type="xsd:string" minOccurs="0"/>

                        <xsd:element name="City" type="xsd:string"/>

                        <xsd:element name="ZIPCode" type="xsd:string" minOccurs="0"/>

                        <xsd:choice>

                              <xsd:element name="State" type="xsd:string"/>

                              <xsd:element name="Country" type="xsd:string"/>

                        </xsd:choice>

            </xsd:sequence>

      </xsd:complexType>

</xsd:schema>

<xsd:schema xmlns:xsd="https://www.w3.org/2001/XMLSchema"

targetNamespace="urn:letter:tags"

elementFormDefault="qualified" xmlns="urn:letter:tags">

      <xsd:element name="person" type="xsd:string"/>

      <xsd:element name="date" type="xsd:date"/>

      <xsd:element name="title" type="xsd:string"/>

      <xsd:element name="URI" type="xsd:anyURI"/>

      <xsd:element name="email" type="emailType"/>

      <xsd:simpleType name="emailType">

            <xsd:restriction base="xsd:string">

                  <xsd:pattern value="[\w\-\.]+@([\w\-]+\.)+[\w\-]{2,4}"/>

            </xsd:restriction>

      </xsd:simpleType>

</xsd:schema>'

go

Now let’s create a table with a typed XML column

CREATE TABLE T (iCol int primary key, xmlCol XML(SC_Letter))

go

And finally, let's insert a sample XML instance

INSERT INTO T VALUES (1, '

<l:letter xmlns:l="urn:letter" xmlns:t="urn:letter:tags">

      <l:date>2006-10-15-08:00</l:date>

      <l:sender>

            <l:FirstName>James</l:FirstName>

            <l:LastName>Kirk</l:LastName>

            <l:Line1>1234 Elm Street</l:Line1>

            <l:City>Redmond</l:City>

            <l:ZIPCode>98052</l:ZIPCode>

            <l:State>WA</l:State>

      </l:sender>

      <l:recipient>

            <l:FirstName>Jean-Luc</l:FirstName>

            <l:LastName>Picard</l:LastName>

            <l:Line1>13 rue des Lilas</l:Line1>

            <l:City>Ronchin</l:City>

            <l:ZIPCode>59555</l:ZIPCode>

            <l:Country>France</l:Country>

      </l:recipient>   

Dear Sir,

      I received your letter dated <t:date>2006-10-10+00:00</t:date> this morning. It would be my pleasure to attend the annual symposium <t:title>XML database systems in class-B Federation Spacecrafts</t:title>.

Please contact my assistant <t:person>Mr. Spock</t:person> at <t:email>spock@space-federation.org</t:email> to make all necessary arrangements.

Sincerely,

James T. Kirk

</l:letter>')

go

As you can see the 'letter" element contains interleaved element and text children. This is legal because the type was declared with mixed=''true".

It is important to understand that this property is not recursive. In our example it would be illegal to add text nodes under the "recipient" element since its type is not mixed. You can for example try the following INSERT query.

INSERT INTO T VALUES (0, '

<l:letter xmlns:l="urn:letter" xmlns:t="urn:letter:tags">

      <l:date>2006-10-15-08:00</l:date>

      <l:sender>

            <l:FirstName>James</l:FirstName>

This is a text node

            <l:LastName>Kirk</l:LastName>

            <l:Line1>1234 Elm Street</l:Line1>

            <l:City>Redmond</l:City>

            <l:ZIPCode>98052</l:ZIPCode>

            <l:State>WA</l:State>

      </l:sender>

</l:letter>')

go

There is a text node under the “sender” element. Since “sender” is of type “addressType” and that type wasn’t defined as mixed, its presence violates the schema and the query predictably fails with the following error message:

XML Validation: Text node is not allowed at this location, the type was defined with element only content or with simple content. Location: /*:letter[1]/*:sender[1]

Next time we’ll look at text nodes, how the validator handles them, and how we can query them…

First let’s create a schema collection with complex types and elements.

CREATE XML SCHEMA COLLECTION SCnil_cplx AS '

<xs:schema xmlns:xs="https://www.w3.org/2001/XMLSchema" targetNamespace="urn:nil" xmlns="urn:nil">

      <xs:complexType name="CT_emptiable">

            <xs:choice minOccurs="0" maxOccurs="unbounded">

                  <xs:element name="a" type="xs:string"/>

                  <xs:element name="b" type="xs:byte"/>

            </xs:choice>

      </xs:complexType>

      <xs:complexType name="CT_nonEmptiable">

            <xs:choice minOccurs="1" maxOccurs="unbounded">

                  <xs:element name="a" type="xs:string"/>

                  <xs:element name="b" type="xs:byte"/>

            </xs:choice>

      </xs:complexType>

      <xs:element name="N1" type="CT_emptiable" nillable="true"/>

      <xs:element name="E1" type="CT_emptiable" nillable="false"/>

      <xs:element name="N2" type="CT_nonEmptiable" nillable="true"/>

      <xs:element name="E2" type="CT_nonEmptiable" nillable="false"/>

</xs:schema>'

go

CREATE TABLE T (idCol int primary key, xmlCol XML(SCnil_cplx))

go

Let’s insert 4 rows, containing instances of the 4 top level elements defined in the schema collection.

INSERT INTO Tcplx VALUES (1,'<ns:N1 xmlns:ns="urn:nil"><a>Data</a><b>2</b></ns:N1>')

INSERT INTO Tcplx VALUES (2,'<ns:N2 xmlns:ns="urn:nil"><a>Data</a><b>2</b></ns:N2>')

INSERT INTO Tcplx VALUES (3,'<ns:E1 xmlns:ns="urn:nil"><a>Data</a><b>2</b></ns:E1>')

INSERT INTO Tcplx VALUES (4,'<ns:E2 xmlns:ns="urn:nil"><a>Data</a><b>2</b></ns:E2>')

go

Now let’s see what happens when we delete all the children elements under each of those top level elements.

Let’s start with element ‘E1’ which is not nillable, and whose content model is emptiable (which means an empty element is valid with respect to the schema definition).

UPDATE Tcplx

SET xmlCol.modify('declare namespace ns="urn:nil"; delete /ns:E1[1]/*')

WHERE idCol = 3

go

SELECT xmlCol FROM Tcplx WHERE idCol = 3

go

The DML operation is successful and following its completion the instance looks like this.

<ns:E1 xmlns:ns="urn:nil" />

Let’s do the same thing with element ‘E2’ which is not nillable and whose content model is not emptiable (which means an empty element is invalid with respect to the schema)

UPDATE Tcplx

SET xmlCol.modify('declare namespace ns="urn:nil"; delete /ns:E2[1]/*')

WHERE idCol = 4

go

As could be expected this query fails with the following message:

XML Validation: Invalid content. Expected element(s): a,b. Location: /*:E2[1]

Now that we’ve established the behavior for non nillable elements let’s look at ‘N1’, which is nillable and whose content model is emptiable.

UPDATE Tcplx

SET xmlCol.modify('declare namespace ns="urn:nil"; delete /ns:N1[1]/*')

WHERE idCol = 1

go

SELECT xmlCol FROM Tcplx WHERE idCol = 1

go

The DML operation is successful and the instance now looks like this

<ns:N1 xmlns:ns="urn:nil" xmlns:xsi="https://www.w3.org/2001/XMLSchema-instance" xsi:nil="true" />

Notice that even though, an empty instance of N1 would have been valid with respect to the schema, when all the children element were deleted the server nilled the instance by adding the attribute xsi:nil with a value of “true”.

Running a similar query on an instance of ‘N2’, which is nillable and non emptiable also yields interesting results.

UPDATE Tcplx

SET xmlCol.modify('declare namespace ns="urn:nil"; delete /ns:N2[1]/*')

WHERE idCol = 2

go

SELECT xmlCol FROM Tcplx WHERE idCol = 2

go

Even though the element isn’t emptiable, the query succeeds, and the resulting instance looks like this:

<ns:N2 xmlns:ns="urn:nil" xmlns:xsi="https://www.w3.org/2001/XMLSchema-instance" xsi:nil="true" />

Once again, when all the children elements are deleted, the element is nilled by adding xsi:nil=”true”.

-
Disclaimer:
This posting is provided “AS IS” with no warranties, and confers no rights.
Use of included script samples are subject to the terms specified at https://www.microsoft.com/info/cpyright.htm.

But first let’s do a quick recap. Any element can be made nillable by adding the attribute nillable=”true” to its declaration. Practically it means that this element can take a special value we call “nil”. An instance of this element will be “nilled” if it has no content and contains the boolean attribute nil from namespace https://www.w3.org/2001/XMLSchema-instance (usually associated to the prefix “xsi” but this is not a requirement) with a value of “true”.

As an example, let’s look at the following schema collection.

CREATE XML SCHEMA COLLECTION SCnil AS '

<xs:schema xmlns:xs="https://www.w3.org/2001/XMLSchema" targetNamespace="urn:nil" xmlns="urn:nil">

      <xs:complexType name="CT_emptiable">

            <xs:choice minOccurs="0" maxOccurs="unbounded">

                  <xs:element name="a" type="xs:string"/>

                  <xs:element name="b" type="xs:byte"/>

            </xs:choice>

      </xs:complexType>

      <xs:simpleType name="ST">

            <xs:restriction base="xs:decimal">

                  <xs:minInclusive value="0"/>

            </xs:restriction>

      </xs:simpleType>

      <xs:element name="E_ST" type="ST" nillable="true"/>

      <xs:element name="E_CT_e" type="CT_emptiable" nillable="true"/>

</xs:schema>

'

go

Now if we create a table with a typed XML column we can easily create and store nilled instances of elements E_ST and E_CT_e

CREATE TABLE T (iCol int primary key, xmlCol XML(SCnil))

go

INSERT INTO T VALUES (1,

'<E_ST xmlns="urn:nil" xmlns:xsi="https://www.w3.org/2001/XMLSchema-instance" xsi:nil="true"/>')

go

INSERT INTO T VALUES (2,

'<E_CT_e xmlns="urn:nil" xmlns:z="https://www.w3.org/2001/XMLSchema-instance" z:nil="true"/>')

go

You may notice that in the second example I used prefix “z” for the https://www.w3.org/2001/XMLSchema-instance namespace. As I said earlier “xsi” is the commonly used prefix but it is by no means the required one.

Now let’s look what being “nilled” means for a simply typed element.

If you query the data of a nilled simply typed element, the result will be the empty sequence. You can verify this by running the following query

SELECT xmlCol.query('declare namespace ns="urn:nil"; data(/ns:E_ST[1])instance of empty()')

FROM T WHERE iCol = 1

The result is “true”.

Now how do we “un-nil” such an element? We could try to delete the xsi”nil attribute but this is not allowed. A query such as

UPDATE T SET xmlCol.modify('declare namespace xsi="https://www.w3.org/2001/XMLSchema-instance"; delete /*[1]/@xsi:nil')

WHERE iCol = 1

will return the following error message

XQuery [T.xmlCol.modify()]: The XQuery syntax '@{https://www.w3.org/2001/XMLSchema-instance}:nil' is not supported.

The solution is to update the value of the element, like this

UPDATE T SET xmlCol.modify('declare namespace ns="urn:nil"; replace value of /ns:E_ST[1] with 1.0 cast as ns:ST?')

WHERE iCol = 1

go

SELECT xmlCol FROM T WHERE iCol = 1

go

The XML instance now looks like this

<E_ST xmlns="urn:nil" xmlns:xsi="https://www.w3.org/2001/XMLSchema-instance">1</E_ST>

Notice that the xsi:nil attribute is no longer present.

The reverse operation consists in updating the element’s value with the empty sequence, like this

UPDATE T SET xmlCol.modify('declare namespace ns="urn:nil"; replace value of /ns:E_ST[1] with ()')

WHERE iCol = 1

go

SELECT xmlCol FROM T WHERE iCol = 1

go

The XML instance now looks like this

<E_ST xmlns="urn:nil" xmlns:xsi="https://www.w3.org/2001/XMLSchema-instance" xsi:nil="true" />

The element’s content is gone and attribute xsi:nil has been added with the Boolean value “true”.

I’m going to stop here for today. In the next installment we’ll look at what happens with complex content elements.

-
Disclaimer:
This posting is provided “AS IS” with no warranties, and confers no rights.
Use of included script samples are subject to the terms specified at https://www.microsoft.com/info/cpyright.htm.

An experienced XQuery user would certainly know when to use the brackets vs. the parentheses, but if you're a new to this technology and are using the article as a starting point, you should be aware that there is a difference. In any case, if you have any questions don't hesitate to use the contact form to send me an email.