Home Page › Forums › BizTalk 2004 – BizTalk 2010 › Order of looping functoid output
- This topic has 1 reply, 1 voice, and was last updated 9 years, 10 months ago by
community-content.
-
AuthorPosts
-
-
April 7, 2009 at 5:43 AM #22097
Hello all,
I got a question about the looping functoid. Is there a way to determine the order of its output?
Let me explain with an example. Consider this schema:
<root>
<record>
<index />
<typeA>
<description />
</typeA>
<typeB>
<description />
</typeB>
<typeC>
<description />
</typeC>
</record>
</root>Where both ‘record’ as all types are unbounded in Max Occurs.
Let’s say I’ve got this input file:
<root>
<record>
<index>1</index>
<typeA>
<description>type A</description>
</typeA>
<typeB>
<description>type B</description>
</typeB>
<typeC>
<description>type C</description>
</typeC>
</record>
<record>
<index>2</index>
<typeB>
<description>again type B</description>
</typeB>
<typeC>
<description>again type C</description>
</typeC>
</record>
<record>
<index>3</index>
<typeA>
<description>another type A</description>
</typeA>
<typeA>
<description>last type A</description>
</typeA>
<typeC>
<description>last type C</description>
</typeC>
</record>
</root>For each type I want a record in the output file like this:
<root>
…
<record>
<index>index node of parent record</index>
<type>name of type (node name)</type>
<description>description of type</description>
</record>
…
</root>So I use a looping functoid for every type in the input file (one functoid and I connect all types to it) and I expect the following output:
<root>
<record>
<index>1</index>
<type>typeA</type>
<description>type A</description>
</record>
<record>
<index>1</index>
<type>typeB</type>
<description>type B</description>
</record>
<record>
<index>1</index>
<type>typeC</type>
<description>type C</description>
</record>
<record>
<index>2</index>
<type>typeB</type>
<description>again type B</description>
</record>
<record>
<index>2</index>
<type>typeC</type>
<description>again type C</description>
</record>
<record>
<index>3</index>
<type>typeA</type>
<description>another type A</description>
</record>
<record>
<index>3</index>
<type>typeA</type>
<description>last type A</description>
</record>
<record>
<index>3</index>
<type>typeC</type>
<description>last type C</description>
</record>
</root>But what I’m getting is:
<root>
<record>
<index>1</index>
<type>typeA</type>
<description>type A</description>
</record>
<record>
<index>3</index>
<type>typeA</type>
<description>another type A</description>
</record>
<record>
<index>3</index>
<type>typeA</type>
<description>last type A</description>
</record>
<record>
<index>1</index>
<type>typeB</type>
<description>type B</description>
</record>
<record>
<index>2</index>
<type>typeB</type>
<description>again type B</description>
</record>
<record>
<index>1</index>
<type>typeC</type>
<description>type C</description>
</record>
<record>
<index>2</index>
<type>typeC</type>
<description>again type C</description>
</record>
<record>
<index>3</index>
<type>typeC</type>
<description>last type C</description>
</record>
</root>So the output is ordered by the type. I’ve noticed that it’s determined by the order of the looping functoid input parameters. Is this example I connected the types to the looping functoid in this order: typeA, typeB, typeC. So when I change it to: typeB, typeA, typeC, all typeB’s are first put in the output, then the typeA’s and then the typeC’s.
I hope someone can tell me if what I want is possible (via the mapping GUI).
If anyone’s wondering: The order matters because the output will eventually be a flat file that’s read by a not so flexible system.
-
April 8, 2009 at 1:15 AM #22104
I got it working. It wasn’t that hard to figure out after all.
In the output I added another node like so:
<root>
…
<record>
<part>
<index>3</index>
<type>typeC</type>
<description>last type C</description>
</part>
</record>
…
</root>I connect the <record> from the input file via a looping functoid to the <record> from the output file. All the type nodes are all connected via a single looping functoid to the <part> node in the output file.
In the eventual flat file the extra <part> nodes don’t matter.
-
-
AuthorPosts
- The forum ‘BizTalk 2004 – BizTalk 2010’ is closed to new topics and replies.