| Author |
Message
|
| drcas |
 Posted: Tue Aug 31, 2004 9:21 am Post subject: Blob to XML |
 |
|
Apprentice
Joined: 03 May 2002
Posts: 26
|
I would appreciate any input to my interface's requirements.
The input message to the message flow will consist of 3 types of 80 byte fixed width messages. One block 1 message- the header-, n number of block 2 messages- the data- and one block 9 message -the trailor. The flow is required to route the message to receivers and filter out the block 2 messages that the receiver does not require and update the trailor with the block 2 count.
I do not believe I can model this message in CWF as CWF does not support repeating groups ( the cardinality of the block 2 records is n). My idea is to bring the message in as a BLOB and then substring it into an XML format. The XML message would then be routed to the receivers and filtered according to the rules. The XML message would then be transformed back to the BLOB and put to the receiver's queues.
Does this sound reasonable or can someone think of a better approach? Note that the message cannot be broken up into separate block messages and the filter rules will be stored in a database as will the receiver's - not that that matters.
I appreciate any assistance.
-David |
|
| Back to top |
|
 |
| jefflowrey |
| Posted: Tue Aug 31, 2004 9:26 am Post subject: |
|
|
Grand Poobah
Joined: 16 Oct 2002
Posts: 19981
|
You can model tagged fixed width data with TDS, which does support variable repeats.
I have done exactly the kind of thing you are looking to do with TDS, when working with Commonline records (student loan data).
_________________
I am *not* the model of the modern major general. |
|
| Back to top |
|
|
| drcas |
| Posted: Tue Aug 31, 2004 9:35 am Post subject: |
|
|
Apprentice
Joined: 03 May 2002
Posts: 26
|
| I'll look into it. What if the data is not tagged? |
|
| Back to top |
|
|
| jefflowrey |
| Posted: Tue Aug 31, 2004 9:47 am Post subject: |
|
|
Grand Poobah
Joined: 16 Oct 2002
Posts: 19981
|
What makes a block 1 different than a block 2 or a block 9? I assume those terms represent tags of some sort.
If not, you can still do this with TDS in some cases - you just have to have record delimiters (like line endings!) or the records need to start with some logically distinct field types.
Or you can still model it as a choice, and then you have to do some work to resolve the choice.
It's likely possible with your data.
_________________
I am *not* the model of the modern major general. |
|
| Back to top |
|
|
| drcas |
| Posted: Tue Aug 31, 2004 12:26 pm Post subject: |
|
|
Apprentice
Joined: 03 May 2002
Posts: 26
|
| The first character in the record is the block type. It is either '1', '2', or '9'. There are no tags in the data. |
|
| Back to top |
|
|
| jefflowrey |
| Posted: Tue Aug 31, 2004 4:25 pm Post subject: |
|
|
Grand Poobah
Joined: 16 Oct 2002
Posts: 19981
|
| drcas wrote: |
| The first character in the record is the block type. It is either '1', '2', or '9'. There are no tags in the data. |
That's a tag.
_________________
I am *not* the model of the modern major general. |
|
| Back to top |
|
|
|
|
