SQL XML Bulk Load and Namespaces

Apr 052011

Handling namespaces in SQL XML Bulk Load can be a headache as there’s not really much documentation about it. However while laborious, it is basically following the normal XML Schema structure.
I’ll not go deep into the SQL XML Bulk Load syntax in this blog post, if you’re looking for that, I’ll advice you to read my SQL XML Bulk Load – basic XSD syntax or SQL XML Bulk Load – more XSD syntax blog posts before this one.

I’ve made this simplified example to illustrate. Suppose you have an XML file something akin to this:

<?xml version="1.0" encoding="utf-8"?>
<Catalogue xmlns:sdt="urn:oasis:names:specification:ubl:schema:xsd:SpecializedDatatypes-2"
           xmlns="urn:oasis:names:specification:ubl:schema:xsd:Catalogue-2"
           xmlns:cac="urn:oasis:names:specification:ubl:schema:xsd:CommonAggregateComponents-2"
           xmlns:cbc="urn:oasis:names:specification:ubl:schema:xsd:CommonBasicComponents-2"
           xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
           xsi:schemaLocation="urn:oasis:names:specification:ubl:schema:xsd:Catalogue-2 UBL-Catalogue-2.0.xsd">
  <cbc:UBLVersionID>2.0</cbc:UBLVersionID>
   <cac:ValidityPeriod>
    <cbc:StartDate>2011-03-02</cbc:StartDate>
  </cac:ValidityPeriod>
   <cac:CatalogueLine>
    <cbc:ID>11002</cbc:ID>    
  </cac:CatalogueLine>
   <cac:CatalogueLine>
    <cbc:ID>11003</cbc:ID>
  </cac:CatalogueLine>
</Catalogue>

I have two custom namespaces, and a mixture of the elements within each namespace. This is simplified, but once the logic is known, it is just expanding it with more namespaces, and more mapping.

First we need to make the base XML schema/XSD file. A way to handle namespaces in the XSD file is to use import statements and split each namespace element in to its own separate file. And that technique works for SQL XML Bulk Load as well.
So we’ll make the base XSD like this

<?xml version="1.0" encoding="utf-8"?>
<xs:schema xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"            
           xmlns:cac="urn:oasis:names:specification:ubl:schema:xsd:CommonAggregateComponents-2"
           xmlns:cbc="urn:oasis:names:specification:ubl:schema:xsd:CommonBasicComponents-2"
           xmlns:xs="http://www.w3.org/2001/XMLSchema"
           xmlns:sql="urn:schemas-microsoft-com:mapping-schema">

  <xs:annotation>
    <xs:appinfo>
      <sql:relationship name="SomeName"
                        parent="Import"
                        parent-key="PK"
                        child="Import2"
                        child-key="FK" />
    </xs:appinfo>
  </xs:annotation>

  <xs:import namespace="urn:oasis:names:specification:ubl:schema:xsd:CommonBasicComponents-2" schemaLocation="cbc.xsd" />
  <xs:import namespace="urn:oasis:names:specification:ubl:schema:xsd:CommonAggregateComponents-2" schemaLocation="cac.xsd" />

  <xs:element name="Catalogue" sql:relation="Import">
    <xs:complexType>
      <xs:sequence>
        <xs:element ref="cbc:UBLVersionID" />
        <xs:element ref="cac:ValidityPeriod" />
        <xs:element maxOccurs="unbounded" ref="cac:CatalogueLine" />
      </xs:sequence>
    </xs:complexType>
  </xs:element>
</xs:schema>

Notice that each of the 3 elements simply has a reference to another element. So the XML tag cbc:UBLVersionID will in the XSD simply be element which ref=”cbc:UBLVersionID”, just as cac:ValidityPeriod element is also just a reference.
The key here however is that I’ve imported two namespaces with a schema location

 <xs:import namespace="urn:oasis:names:specification:ubl:schema:xsd:CommonBasicComponents-2" schemaLocation="cbc.xsd" />
<xs:import namespace="urn:oasis:names:specification:ubl:schema:xsd:CommonAggregateComponents-2" schemaLocation="cac.xsd" />

These lines will tell the base schema file where to look for the specific definitions for the namespaces.

Now, I can make a specific schema file for the cbc namespace and the cac namespace.
The cac one looks like

<?xml version="1.0" encoding="utf-8"?>
<xs:schema xmlns:tns="urn:oasis:names:specification:ubl:schema:xsd:CommonAggregateComponents-2"
                attributeFormDefault="unqualified" elementFormDefault="qualified"
                targetNamespace="urn:oasis:names:specification:ubl:schema:xsd:CommonAggregateComponents-2"
           xmlns:sql="urn:schemas-microsoft-com:mapping-schema"
           xmlns:cbc="urn:oasis:names:specification:ubl:schema:xsd:CommonBasicComponents-2"
           xmlns:xs="http://www.w3.org/2001/XMLSchema">
    <xs:import namespace="urn:oasis:names:specification:ubl:schema:xsd:CommonBasicComponents-2" schemaLocation="cbc.xsd" />

  <xs:element name="ValidityPeriod" sql:is-constant="1" >
    <xs:complexType>
      <xs:sequence>
        <xs:element ref="cbc:StartDate" />
      </xs:sequence>
    </xs:complexType>
  </xs:element>
  <xs:element name="CatalogueLine" sql:relationship="SomeName" sql:relation="Import2">
    <xs:complexType>
      <xs:sequence>
        <xs:element ref="cbc:ID"  />
      </xs:sequence>
    </xs:complexType>
  </xs:element>
</xs:schema>

Here the same principle of using import statement to map a namespace

<xs:import namespace="urn:oasis:names:specification:ubl:schema:xsd:CommonBasicComponents-2" schemaLocation="cbc.xsd" />

is used similar to the base file.

The cac file is used to define the contents of all cac namespace elements (the naming of the file is irrelevant, I simply named them this way to help myself keeping track). And where needed in this cac file, a reference is made to an element in the cbc file (in this case cbc:StartDate and cbc:ID).
The cbc file will then qualify all cbc elements

<?xml version="1.0" encoding="utf-8"?>
<xs:schema xmlns:tns="urn:oasis:names:specification:ubl:schema:xsd:CommonBasicComponents-2"
           attributeFormDefault="unqualified" elementFormDefault="qualified" targetNamespace="urn:oasis:names:specification:ubl:schema:xsd:CommonBasicComponents-2"
           xmlns:xs="http://www.w3.org/2001/XMLSchema"
           xmlns:sql="urn:schemas-microsoft-com:mapping-schema"

           >
  <xs:element name="UBLVersionID" type="xs:decimal" sql:field="VersionID" />
  <xs:element name="StartDate" type="xs:date" sql:field="DateX" />
  <xs:element name="ID" type="xs:unsignedShort" sql:field="LineID" />
</xs:schema>

As can be seen, the sql:field, sql:relationship and sql:relation are all used across all 3 files.
In the base file, I’ve defined a relationship, and I’m using the relationship in the cac file around a cbc element.

Running this through a SQL XML Bulk Load, and I get my data imported into two tables like this

This was just a simplified example, but I hope it helps to show how the syntax is build up in a XSD and then spice it up with some SQL mapping.

For more insight into SQL XML Bulk Load and the syntax, you can check out my other posts on the subject.

COUNT on LEFT JOIN

Mar 022011

When using COUNT with LEFT JOINs, it is worth noting that is that it does not count NULL values if counting over something different than *.
This is useful if you wish to count how many of the rows fulfill some sort of join clause.

Take the following example.
I’ve created two tables consisting of a unique identifier and a text column. I’ve added test data and for every two row I placed into Tabel_1, I placed one row into Table_2. Then I made a left join query which looks like:

SELECT *
FROM Table_1 AS T
LEFT JOIN Table_2 AS T2 ON T.PK = T2.FK

Which provide an output like this:

LEFT JOIN example for using COUNT | EXEC(@sql)

LEFT JOIN example for using COUNT

Now to illustrate the difference in count, run a query like this:

SELECT COUNT(*) AS AllRows, COUNT(PK) AS RowsFromLeft, COUNT(FK) AS RowsFromRight
FROM Table_1 AS T
LEFT JOIN Table_2 AS T2 ON T.PK = T2.FK

Which provide the following output:

COUNT result for LEFT JOIN | EXEC(@sql)

COUNT result for LEFT JOIN

I’ve seen many more or less creative solutions in my time to how to count over the Right table in a Left Join, because few people seem to read the documentation on something as simple as a COUNT.
However often it is well worth it just to take a peek.

The select list for the INSERT statement contains fewer items than the insert list. The number of SELECT values must match the number of INSERT columns

Feb 232011

I ran into the error: The select list for the INSERT statement contains fewer items than the insert list. The number of SELECT values must match the number of INSERT columns.
Well I’ve run into it often.

Normally this simply means you have specified fewer items in values for an INSERT statement than you have specified you need. For example:

INSERT INTO Table1
( ID, [Column], Column1 )
SELECT 1,
'Value',

Now, that’s easily rectified, by including the columns you need or removing the one from INSERT you do not need.

However, I also noticed that in my situation I had all the columns needed, but I was lacking a comma – like this:

INSERT INTO Table1
( ID, [Column], Column1 )
SELECT 1,
'Value1'
'Value2'

Notice the lack of comma between Value1 and Value2.

This gives the mentioned error and not a syntax error in the SELECT statement. The reason is that SQL Server concatenates the Value1 and Value2 fields into one column, and therefore the SELECT only produces 2 columns.
SQL Server will be able to concatenate two columns into one without needing to add them together, but only two columns.
So if I were to forget the third comma I would get an “Incorrect syntax near Value2″ error message alerting me to the issue. But forget one comma between 2 columns and you get no such information.

I didn’t see a mention of this in the MSDN documentation.

Migrating to wordpress

Nov 262010

I’m migrating my blog system to WordPress, so hopefully the blog will be up in full force over the weekend

@@rowcount and print

Jun 072010

Using @@rowcount is nice at times to get the number of rows affected by the last SQL statement. However sometimes you get an unexpected result.
One such situation is with the print statement which is often used as debugging statement in a query.
If looking at the SQL statement:
SELECT 1
UNION
SELECT 2
SELECT @@ROWCOUNT

Then the @@rowcount naturally will return the value 2.
However if we insert a print statement just before the rowcount like this:

SELECT 1
UNION
SELECT 2
PRINT ‘test’
SELECT @@ROWCOUNT

Rowcount now suddenly is 0.

That means the print statement changes the rowcount despite it being just a print to the console window.
Something to be aware of if relying on @@rowcount.

The MERGE syntax - syntax and performance

May 212010

Normally when you have rows from one table you want to move over into the other, you’ll have to run both an update query to get your existing rows updated with the new values and insert query to get new rows over to your target. That means you have to write two queries. With the merge, you only need one as it performs both the update and the insert.
So let’s take a look at it.

First, I’ll create 2 tables with dummy data; a tblSource and a tblTarget.
The scripts for creating the tables and data can be found here merge_setup_20100521.sql (3.58 kb), but basically it’s just a target and source table with an ID, ValA, ValB and ValC with a clustered index on ID and some dummy data.
So to do the update/insert it would look like this:
UPDATE MyTarget SET
MyTarget.ValA = MySource.ValA,
MyTarget.ValB = MySource.ValB,
MyTarget.ValC = MySource.ValC
FROM dbo.tblTargetInUp AS MyTarget
INNER JOIN dbo.tblSourceInUp AS MySource ON MyTarget.ID = MySource.ID

INSERT INTO dbo.tblTargetInUp
SELECT ID, ValA, ValB, ValC
FROM dbo.tblSourceInUp AS MySource
WHERE NOT EXISTS (SELECT 1 FROM dbo.tblTargetInUp AS MyTarget WHERE MyTarget.ID = MySource.ID)

If we look at the execution plan, I get an estimated cost of 9.1754 for the update and 2.48725 for the insert, meaning a combined cost of 11.66265 for 33.333 rows updated and 33.3334 rows inserted, into a target of 66.666 rows.

Now do remember, the estimated cost is just a number for how the query runs in my environment, it can’t be taken as a direct number and transferred to another system – I’m only interested in the relative comparison with the merge. More data, more indexes, more variations will all affect the actual numbers. Also if looking at the execution plan, it is clear it is two queries we fire, meaning that all the overhead which goes into running one query will be doubled for this. But it was how you’d have to do merges in the past.

Now, with the merge syntax we can do it like this:
MERGE dbo.TblTargetMerge AS MyTarget USING
(
SELECT * FROM dbo.TblSourceMerge AS TS
) AS MySource
ON MyTarget.ID = MySource.ID
WHEN MATCHED THEN
UPDATE SET
MyTarget.ValA = MySource.ValA,
MyTarget.ValB = MySource.ValB,
MyTarget.ValC = MySource.ValC
WHEN NOT MATCHED THEN
INSERT
(ID, ValA, ValB, ValC)
VALUES
(MySource.ID, MySource.ValA, MySource.ValB, MySource.ValC)
;

Note that the syntax takes both the update and insert in the WHEN MATCHED and WHEN NOT MATCHED.
For the complete overview over the syntax, I’ll refer you to the documentation by microsoft: MERGE (Transact-SQL).
But basically – you MERGE into a table using a source, and then define the ON clause (as you would a join), and then specify the WHEN MATCHED and the WHEN NOT MATCHED clauses.

One thing I’ll expand on myself though, is the OUTPUT clause which can also be coupled on to the merge. I mentioned the OUTPUT clause myself recently.
The important thing is that you can couple the $action to the output clause and get information about whether you merged the data or you inserted the data, meaning whether the row was matched or not matched.
Like this:

MERGE dbo.TblTargetMerge AS MyTarget USING
<…snipped FOR being brief…>
OUTPUT $ACTION, rest_of_select_here
;

Nifty.
Anyways – once we’ve build this query, we can look at the execution plan, and here we clearly see it is handled as one query.
And in my case, the estimated cost is 8.407 for the same number of rows as above, meaning we’ve saved about 27.9% just by changing syntax.

Now, this structure I used to compare is very simple – with a simple matching of ID to ID and then just insert/overwrite everything so the actual result may vary (naturally), however the merge syntax does appear to be faster and with the added bonus of keeping the query combined into one syntax rather than divided into two different queries.

I must admit, I do like the merge syntax myself once I learned to read and write them.

A STR issue

May 202010

I’ve seen many uses, and some misuses, of the STR function over time.

One of the more “easy to spot problems” is that the STR function returns char datatype, meaning it’ll pad the result with spaces and people need to trim the result.

Today I saw an even worse issue. The STR function takes a float argument. This means if you feed it a string (yes, I’ve seen it done) it will implicit convert that string to a float if it can. That in itself can cause an error, but even worse, it can cause a difficult to find bug.
Suppose you have STR(’0001′) and run that, you’ll after trimming end up with the result of ’1′. Why?
Because ’0001′ is converted to a float, which – as we all know – is 1, which then will be cast to a string.

Changing domain

May 192010

I’ve decided to change my previous domain “www.sqlstuff.dk” to this new one “www.execsql.com” – I apologize for any inconvenience which might happen as a result, but it is better done now then later :)

All access to sqlstuff should get a 301 redirect to the new www.execsql.com domain and hopefully the transition should be smooth.

So welcome to EXEC(@sql) and bye bye to SQLStuff.dk :)

 

 

SQL XML Bulk Load - more XSD syntax

Apr 172010

I have posted a couple of times about the XML/XSD syntax used for SQL XML Bulk Load but mostly the syntax have been pretty basic, so I wanted to expand a little on the syntax with a more complex example.

So this time I’ll have an XML file structured like this. As can be seen this XML file consists of a couple of nested tags and is a structure that’s realistic when it comes to data imports. In fact it is a minimization of an existing one I have used for work, where I have removed duplicated structures to provide a cleaner example.
The file consists of repeated details tags with content and attributes used, a product files section containing a repeated tag and a nesting of Product Lists. It should cover most of the situations which one would encounter.
The only thing I do not use in this instance is composite keys, but they’re no different compared to using a singular one. I’ve written about composite keys and SQL XML Bulk Load here.
Anyways, let’s tackle the XSD for this file. (If new to the SQL XML Bulk Load, please read my SQL XML Bulk Load – basic XSD syntax post first)

The first tag, ProductID, is simple enough and covered in the basic syntax blog post.
It requires a database table containing this field. Any subsequent values placed on the same level of the XML file would be included into the same table.
So we’ll make a Product table:
CREATE TABLE Product(
ProductID INT NOT NULL
)

The XSD field will then be mapped to this table. And because I’ve named the field the same as the tag in the XML file I do not need to map the field.
This gives me the segment:
<xs:element Name=“Product” sql:relation=“Product” maxOccurs=“unbounded”>
<xs:complexType>
<xs:sequence>
<xs:element Name=“ProductID” />

The details section is next. For this we need to set up a relationship because the section can be repeated several times, so we need to map it out into a related table.
For this section we’ll need the table:
CREATE TABLE Details(
Name VARCHAR(MAX) NULL,
[Description] VARCHAR(MAX) NULL,
LanguageCode VARCHAR(5) NOT NULL,
ProductID INT NOT NULL
)

This can then hold the Name and Description tags from the section, and the attribute “LanguageCode” and the key ProductID to map it to the Product table/Product section.
Because we need a relationship we’ll have to map it out:
<sql:relationship Name=“Detail”
parent=“Product”
parent-key=“ProductID”
child=“Details”
child-key=“ProductID”
/>

And this then provides the following section or the XSD
<xs:element Name=“Details” sql:relation=“Details” sql:relationship=“Detail”>
<xs:complexType>
<xs:sequence>
<xs:element Name=“Name” />
<xs:element Name=“Description” />
</xs:sequence>
<xs:attribute Name=“LanguageCode” Type=“xs:string” use=“required” />
</xs:complexType>
</xs:element>

As can be seen we map the section to the relationship and the table relation.
For the section of Files, we’ll need to expand a little more. Again we need a relationship to tell the bulk load which keys to use.
First the table to import into:
CREATE TABLE Files (
ProductID INT NOT NULL,
[Type] VARCHAR(255) NULL,
[FileName] VARCHAR(255) NULL
)
Then the relationship in the XSD file
<sql:relationship Name=“ProductFiles”
parent=“Product”
parent-key=“ProductID”
child=“Files”
child-key=“ProductID”
/>

Just like the previous one.

The XSD section will then look like the following:
<xs:element Name=“ProductFiles” sql:is-constant=“1″>
<xs:complexType>
<xs:sequence>
<xs:element maxOccurs=“unbounded” Name=“FileName”
sql:relationship=“ProductFiles” sql:relation=“Files”>
<xs:complexType>
<xs:simpleContent>
<xs:extension base=“xs:string”>
<xs:attribute Name=“type” Type=“xs:string” use=“required” />
</xs:extension>
</xs:simpleContent>
</xs:complexType>
</xs:element>
</xs:sequence>
</xs:complexType>
</xs:element>

We have the is-constant=”1” because this section can only exists once. The relationship and relation is put on the “FileName” element. The “tricky” part here is the xs:extension which is needed because otherwise the Bulk Load will throw an error that you’re trying to map multiple FileName elements. <ERROR>

On to the List section.
Now this is a more difficult one, because here we’ll need a double relationship. We need to map the ProductID to the Type and then the Type to the multiple ProductID’s within that section. So we need the following tables:
CREATE TABLE ProductList (
[ID] INT IDENTITY(1,1) NOT NULL,
ProductID INT NOT NULL,
Header VARCHAR(255) NULL
)
CREATE TABLE ProductListProduct(
ProductListID INT NOT NULL,
ProductID INT
)

Here we make an Identity ID in the ProductList which will be the foreign key in the ProductListProduct:ProductListID column.
So we need the two following relationships in our XSD
<sql:relationship Name=“ProductToList”
parent=“Product”
parent-key=“ProductID”
child=“ProductList”
child-key=“ProductID”
/>

<sql:relationship Name=“ListToProduct”
parent=“ProductList”
parent-key=“ID”
child=“ProductListProduct”
child-key=“ProductListID”
/>

To map the relationship. The only difference here is that the parent-key for ListToProduct is the identity column from the ProductList table, and it will be used as the child-key in the column ProductListID.
And this gives the following XSD section

<xs:element Name=“ProductList” sql:is-constant=“1″>
<xs:complexType>
<xs:sequence>
<xs:element Name=“List” sql:relationship=“ProductToList” sql:relation=“ProductList”>
<xs:complexType>
<xs:sequence>
<xs:element maxOccurs=“unbounded” Name=“ProductID” Type=“xs:integer”
sql:relationship=“ListToProduct” sql:relation=“ProductListProduct” />
</xs:sequence>
<xs:attribute Name=“header” Type=“xs:string” use=“required” />
</xs:complexType>
</xs:element>
</xs:sequence>
</xs:complexType>
</xs:element>

Again we have the is-constant annotation.
We then map the element List to the relationship ProductToList as defined previously and to the table ProductList. Because header is then an attribute named the same as the field in the database, it’ll map automatically as usual.
However we’ll then map the inner element “ProductID” within the List element to the next relationship ListToProduct and the table ProductListProduct.

The SQL XML Bulk Load will then automatically handle the identity column in the ProductList table and map it to the ProductListID column in the ProductListProduct table.

This gives us this following complete XSD file

And when running the bulk load with this XML and this XSD we get the following result:

Result for importing SQL XML Bulk Load

SQL XML Bulk Load, Relationships and composite keys

Apr 142010

A little quick note, when needing composite keys in your relationship in SQL XML Bulk Load the syntax is as follow:

<sql:relationship name=“Detail”
parent=“Import.Product”
parent-KEY=“ProductID FileID”
child=“Import.ProductDetails”
child-KEY=“ProductID FileID”/>

meaning they’re simply space separated in the relationship annotation. The composite key in this aspect are the columns ProductID and FileID.
As can also be seen in this annotation, I’m using a schema qualifiaction with Import, so if you have other schemas, you can simply write them in the parent/child.

For more syntax, please refer to the blog post SQL XML Bulk Load – basic XSD syntax

Older Entries