A more complete solution on how to handle this is by having different names for TAX1 & TAX2. You can accomplish this by specifying an alternate name for the two TAX segments in the flat file schema. If you do this, you’ll have to alter both the inbound (convertToValues) and outbound (convertToString) maps that use this flat file schema. The converToString service would no longer recognize TAX as a valid segment - the segments would need to match the alternate names the users define. The final name in the output record would still be TAX, but the mapping would need to change to reflect the alternate name.
To do this:
Create a new TAX segment in the flat file schema. This cannot be a reference, it must be a definition. (<-- how the @$#!+%! could anyone have known this) Re-create this TAX segment to look just like the definition that is currently in place. Delete the existing reference to TAX record. Do this process again for the second TAX reference. Make sure the schema looks the same as the original schema structure. The only difference should be that you now have two TAX record definitions instead of references. Now specify an alternate name for the first TAX segment (TAX1) and a different alternate name for the second segment (TAX2). Now those TAX segments should always be referred to in mapping as TAX1/TAX2.
And then…it eventually does work.
#edi#Integration-Server-and-ESB#webMethods