Biztalk schema pipe delimited header/detail on same level

Question

I have a flat file below. I'm having trouble with the schema with this layout. I switched it around to have a header and a detail and created an application with no problems, but the customer won't change the layout. This is probably pretty basic, but I'm a beginner. How do I take certain fields from this layout and create a header and a detail? The last date field needs to be in the header so you can see how it's random.

PO207730CO|1271|customer 1|john doe|1|161075|161075|BROOM FLAGGED LOBBY|2|5.62|24-Feb-2014|
PO207730CO|1271|customer 1|john doe|2|167316|167316|CLEANER DISPATCH SPRAY HOSPITAL DISINFECTANT W/BLEACH|1|59.84|24-Feb-2014|
PO207730CO|1271|customer 1|john doe|3|162175|162175|DUST PAN LOBBY|2|6.26|24-Feb-2014|
PO207730CO|1271|customer 1|john doe|4|163325|163325|MOP WET LARGE GENERAL-PURPOSE BLUE WB/LP|1|18.45|24-Feb-2014|
PO207730CO|1271|customer 1|john doe|5|164715|164715|SOAP PROVON MEDICATED TFX|1|32.79|24-Feb-2014|
PO207730CO|1271|customer 1|john doe|6|166338|166338|TOWEL MULTI-FOLD SCOTT WHITE|5|18.91|24-Feb-2014|
PO207814CO|1264|customer 2|jane doe|1|Cups||Bib 20x35 2 Ply Lab (756220)|1|17.47|24-Feb-2014|
PO207814CO|1264|customer 2|jane doe|2|Cups||Cup 9oz Translucent (098219)|1|24-Feb-2014|
PO207814CO|1264|customer 2|jane doe|3|Cups||Cup Foam 16oz (177190)|2|35.1|24-Feb-2014|
PO207814CO|1264|customer 2|jane doe|4|Cups||Lid 16/20 Whte Tab W/Sslot (194088)|2|16.57|24-Feb-2014|
PO207814CO|1264|customer 2|jane doe|5|Cups||Tissue 2-Ply 100-Sht (343227)|3|16.38|24-Feb-2014|

Answer 1

Thanks for the replies.

I went with an SSIS package to preload a table and build the pipe delimited file. I have Biztalk then picking up that file and debatching it the way I need it.

Answer 2

The basic problem here is the Flat File Disassembler does not support the concept of splitting/debatching based on changing values, PO207730CO -> PO207814CO for example.

So, you'll have to regroup by PO number at some following step.

You have a few options:

1. Use a Custom XSLT Map to group the lines based on the PO Number, then split, by using a Receive Pipeline on an Orchestration for example. https://social.technet.microsoft.com/wiki/contents/articles/17985.xslt-muenchian-grouping-biztalk-complex-transformation.aspx
2. Use an xPath Debatching Pattern in an Orchestration. http://www.biztalkgurus.com/biztalk_server/biztalk_2004/m/biztalk_2004_samples/32438.aspx

Either way, you would parse the flat file as you are now, row by row.