Ureader.com  
Microsoft software help and Community
   home   |   control panel login   |   archive   |  
 
Biztalk
acce.forsuppliers
acce.hipaa
acce.hl7
acce.rosettanet
accelerator.finserv
admin
appintegration
framework
general
library
nonxml
orchestration
sdk
server
setup
tools
xlangs
xsharp
  
 
date: 30 Mar 2006 11:12:48 -0800,    group: microsoft.public.biztalk.nonxml        back       


Flat File Disassembler   
Hi,
I have a flat file with a basic structure:

PONumber|LineNumber|LineDetails

A file contains many of these lines, including several POs with several
lines each.

What I want the disassember to do, is to separate the flat file into
envelopes by PONumber.

So, basically, every time it sees a new PONumber, to start a new xml
message.

How do I do this?

Can it be done in a disassembler? Custom pipeline component?

I know I could break all the lines up and then use a correlation to put
the POs back together, but it would be better if I could do this in the
pipeline.

Any ideas are appreciated.

Ana
date: 30 Mar 2006 11:12:48 -0800   author:   AnaB

Re: Flat File Disassembler   
The standard Flat File Disassembler cannot split a message based on data as 
you wish.
The normal way of handling this situation is to use a map, first disassemble 
the flat file and then map, grouping the records by PONumber.
The map for this uses the <xsl:key> and requires a custom Xslt in the map.

Alternatively you can write your own FlatFile Disassembler.

Greg


"AnaB"  wrote in message 
news:1143745968.597890.14460@j33g2000cwa.googlegroups.com...
> Hi,
> I have a flat file with a basic structure:
>
> PONumber|LineNumber|LineDetails
>
> A file contains many of these lines, including several POs with several
> lines each.
>
> What I want the disassember to do, is to separate the flat file into
> envelopes by PONumber.
>
> So, basically, every time it sees a new PONumber, to start a new xml
> message.
>
> How do I do this?
>
> Can it be done in a disassembler? Custom pipeline component?
>
> I know I could break all the lines up and then use a correlation to put
> the POs back together, but it would be better if I could do this in the
> pipeline.
>
> Any ideas are appreciated.
>
> Ana
>
date: Fri, 31 Mar 2006 08:25:38 +1200   author:   Greg Forsythe

Google
 
Web ureader.com


    COPYRIGHT 2007, YARDI TECHNOLOGY LIMITED, ALL RIGHT RESERVE  |   contact us