Support different validations for the same segment at different levels about staedi HOT 12 CLOSED

@benisantos - this is supported using "implementation" schemas. Take a look at the Validation page on the wiki. Basically, you define the transaction schema as the base standard, then you customize the segments using <implementation>. If it ends up being too large, you can always defined the implementation in a separate file and include the other schema that defines the transaction and all the segment and element types.

That example in the wiki is very basic (just a single segment), but the idea applies to both segments and loops. Let me know if you have further questions. The example could probably be a bit more complex.

from staedi.

I'm taking a look at src/test/resources, but I have a doubt:

• There are examples of reusing a transaction segment in the implementation without changes. For instance:

  <segment type="AK2">
    <!-- Unchanged from standard -->
  </segment>

• But I don't find an example of reuse a full loop in the same way. Is supported something like that on an implementation?

  <loop code="foo" type="bar">
    <!-- Unchanged from standard -->
  </loop>

from staedi.

@benisantos - that is not supported. Typically, implementations of a loop will reduce the occurrences of segments or remove them altogether, so I don't think there would really be a case for using the loop as-is. In the case of PNRGOV, the document you linked is the implementation guide. I have not been able to locate the standard definition, but I believe it would likely be a more general description of the message. For example, all references to TVL would be identical. It's only in the implementation guide where the different TVL requirements are specified.

Do you know if the PNRGOV standard is available anywhere?

from staedi.

@MikeEdgar I don't know any standard documentation like you suggest :(
I'll try your transaction-vs-implementation advice as soon as I can.
Anyway, Thank you for your help!

from staedi.

@MikeEdgar I think I found an error related to transaction-vs-implementation schemas and empty segments (like SRC in PNRGOV), when the empty segment is the first element of the loop (in other positions works perfectly).

If I change the transaction and implementation schemas to remove SRC and remove as well from the EDIFACT file, the validation is OK. Same, if I put SRC as second or third position of the loop. But in first position, the validation process fails.

Could you realize where the problem is? Or do you need a failed test to help you?

from staedi.

@benisantos , I will take a look later today, but a test case would definitely be appreciated!

from staedi.

@MikeEdgar I have created a custom issue (#79) for the problem with empty segments starting loops. And I have attached a failed test case :)

from staedi.

@benisantos - how are things going with your implementation using the PNRGOV guide? I wanted to check if there is anything still needed on this issue.

from staedi.

Sorry @MikeEdgar , I'm busy with other business now :( I'll go back to this issue as soon as I can.

from staedi.

Hi @benisantos - I just wanted to follow up on this one to see if you were able to solve your schema issue.

from staedi.

Hi @MikeEdgar ! Yes, the schema issue was solved with the advice of question #104. This question could be closed :)

And thank you for this awesome lib!

from staedi.

@benisantos , thank you, glad to hear that!

from staedi.