Grouping in the destination schema

good day Gopi,

this would be trivial in Mercator, but it is more challenging in Biztalk mapper.

what you can do is use a scriptoid attached to a cumulative to read in all the order IDs and store them in a global dictionary as keys. this will give you a hash of all orders that can be used as a set of unique order ids.

then you can have another scriptoid output each unique order ID and use it in a logical tests in an evaluation where you loop through all orders only meeting a true case when you have a match with the provided unique order number.

it is difficult to explain, but you might be able to phanaggle it to work. To reduce the algorithm, here it is

– get all unique order ids

– drive shipments output loop with orders input loop

– only allow shimenetorder to be created when

    – there are more unique order IDs left

For each shimenet order, you are looping through all shipments and

     – only allow package id to be created when there is a match between the order id in the source shipment and the scriptoid-provided unique order id

It will take an extreme amount of experience with the mapper to pull this off.  Maybe someone else has a better idea. This example illustrates one of the things about Biztalk mapper that makes me uncomfortable – it forces you to drive loop iterations off of "features of the input data" rather than for business reasons.  Either that or write a .net app to pre-sort the data before you map it – but that would just be an admission of BizTalk being a poor mapping tool, wouldn't it?

 – wa