Splitting Messages

Many messages passing through Interlok will consist of multiple elements. Often you will get a batch of messages (popular if you’re a heavy EDI user); but each message needs to be handled individually. This is where the basic-message-splitter-service and advanced-message-splitter-service are useful.

The key difference basic-message-splitter-service and advanced-message-splitter-service is that advanced-message-splitter-service allows you to apply an arbitrary set of services to each split message; whereas basic-message-splitter-service does not. This means that advanced-message-splitter-service is able to handle split messages inline; basic-message-splitter-service forces you to use an external messaging system like JMS to handle inter-workflow communications.

Basic Message Splitting

<basic-message-splitter-service>
  <connection class="jms-connection">
 ... Configuration skipped
  </connection>
  <producer class="jms-topic-producer">
    <destination class="configured-produce-destination">
      <destination>split.message</destination>
    </destination>
  </producer>
  <splitter class="line-count-splitter">
    <split-on-line>10</split-on-line>
  </splitter>
</basic-message-splitter-service>

• Uses a line-count-splitter to create a new message every 10 lines.
• Each split message is published to the JMS topic split.message

Advanced Message Splitting

Rather than containing a producer and connection combination, the advanced-message-splitter-service wraps a Service to define the functionality required for each split message. Again, we are using standard components to compose the behaviour that is required.

To replicate the same behaviour as basic-message-splitter-service using advanced-message-splitter-service then we could have something like this:

<advanced-message-splitter-service>
 <splitter class="line-count-splitter">
  <split-on-line>10</split-on-line>
 </splitter>
 <service class="service-list">
  <services>
   <standalone-producer>
    <connection class="jms-connection">
   ... Configuration skipped
    </connection>
    <producer class="jms-topic-producer">
      <destination class="configured-produce-destination">
        <destination>split.message</destination>
      </destination>
    </producer>
   </standalone-producer>
  </services>
 </service>
</advanced-message-splitter-service>

• Uses a line-count-splitter to create a new message every 10 lines.
• Each split message is published to the JMS topic split.message

Splitter Implementations

The standard implementations of MessageSplitter are:

Splitter Examples

It’s probably best to describe behaviour using an example. You’ll be able figure out the best splitter for your use-case; in the following sections the payload is shown first (where applicable) then an XML configuration snippet and finally the resulting behaviour.

Line count splitter

First,Second,Team
A1,A2,Team1
B1,B2,Team1
C1,C2,Team1
D1,D2,Team2
E1,E2,Team3
F1,F2,Team3

<splitter class="line-count-splitter">
  <split-on-line>1</split-on-line>
  <keep-header-lines>1</keep-header-lines>
</splitter>

• line-count-splitter creates 6 messages.
  • Each message will have First,Second,Team as the first line.

XPath Message Splitter

<envelope>
  <document>one</document>
  <document>two</document>
  <document>three</document>
</envelope>

<splitter class="xpath-message-splitter">
 <xpath>/envelope/document</xpath>
 <encoding>UTF-8</encoding>
</splitter>

• xpath-message-splitter creates 3 messages, each only containing a <document> element.

Xpath Document Copier

<envelope>
  <document>one</document>
  <document>two</document>
  <document>three</document>
</envelope>"

<splitter class="xpath-document-copier">
  <xpath>count(/envelope/document)</xpath>
</splitter>

• xpath-document-copier creates 3 copies of the original document.

Split By Metadata

The message contains the following metadata:

recipients=Alice,Bob,Carol

<splitter class="split-by-metadata">
 <separator>,</separator>
 <metadata-key>recipients</metadata-key>
 <split-metadata-key>recipient</split-metadata-key>
</splitter>

• split-by-metadata creates 3 new messages, each of which will contain a recipient metadata key.
  • The first will contain recipient=Alice; the second recipient=Bob, and the last recipient=Carol

Mime Part Splitter

Content-Type: multipart/related;
  boundary="_005_AA217B98DF7BEB479FEA1BAF75D31ADBA57C11mtanobodycom_";
  type="multipart/alternative"
MIME-Version: 1.0

--_005_AA217B98DF7BEB479FEA1BAF75D31ADBA57C11mtanobodycom_
Content-Type: text/plain; charset="us-ascii"

The quick brown fox jumps over the lazy dog.

--_005_AA217B98DF7BEB479FEA1BAF75D31ADBA57C11mtanobodycom_
Content-Type: text/plain; charset="us-ascii"

Pack my box with a dozen liquor jugs.

--_005_AA217B98DF7BEB479FEA1BAF75D31ADBA57C11mtanobodycom_
Content-Type: text/plain; charset="us-ascii"

Quick zephyrs blow, vexing daft Jim.

--_005_AA217B98DF7BEB479FEA1BAF75D31ADBA57C11mtanobodycom_
Content-Type: text/plain; charset="us-ascii"

How quickly daft jumping zebras vex.

--_005_AA217B98DF7BEB479FEA1BAF75D31ADBA57C11mtanobodycom_--

<splitter class="mime-part-splitter"/>

• mime-part-splitter creates 4 split messages.

Regexp Message Splitter

This type of splitter can be as complicated or as simple as your regular expression skills; for our purposes let’s say that the document is:

A1,A2,Group1
B1,B2,Group1
C1,C2,Group1
D1,D2,Group2
E1,E2,Group3
F1,F2,Group3

Our configuration is:

<splitter class="simple-regexp-message-splitter">
 <split-pattern>&#x0A;</split-pattern>
 <compare-to-previous-match>true</compare-to-previous-match>
 <ignore-first-sub-message>true</ignore-first-sub-message>
 <match-pattern>^[^,]+,[^,]+,([^,]+)</match-pattern>
</splitter>

• simple-regexp-message-splitter creates 3 messages, grouped by the 3rd field of the CSV message GroupX.