DBL Repository embedded group fields, structures and arrays need proper formatting and handling.

[tscaff]

We are facing a interesting challenge with some fairly complex structures with embedded fields that have group overlays assigned to other structures.

The subfields with groups within groups are not being broken out into true JSON arrays. Once the subfield structure goes beyond the initial first level, it defaults to creating flat string elements without the group field elements broken out.

I would like for CodeGen to support recursive expansion of the groups within groups that are arrays to the proper JSON array structure. The Synergery/DBL language supports this construct, so I feel that the CodeGen and Harmony Core templates should be able to support this brake out also.

See attached for an example where the group structure within a structure was broken out into separate numbered field properties, instead of an embedded array with the proper field properties within those array elements.

We have considered and implemented broken out versions of the embedded group structures and arrays into individually repository structures and added those data elements to ArrayLists on return.

Though not common, these embedded structures, due to the design chosen, are part of an ISAM database file and are stored and utilized in production. So refactoring created additional work to break out the elements in to individual array lists.

[SteveIves]

Hi Tom,

Sorry for the delay in responding to this one. I understand your ask, and it's not an unreasonable one, but it's one that will be very difficult to deliver. The reason for the behavior was that WAY WAY back when CodeGen was first created (circa 2007) the primary use case at the time was generating code for SQL Replication, and in that world there are no overlays, or arrays, or groups, or arrays of groups, or group overlays. There are only columns. So CodeGen was designed to work in that environment, where the only option was to flatten arrays, and groups, and arrays of groups into individual columns.

Clearly things have moved on, and now we have OO, and web services, and data constructs like JSON that deals very nicely with heirarchical data. And you're absolutely right, it would be great to have better support for that in CodeGen. But getting from here to there is going to be a very non-trivial journey.

All I can say for now is that i'll give it some serious thought and see if I can come up with a plan. Part of the issue is that I'm not sure whether repository can adequately describe all of the possible heirarchical relationships, but for sure it can describe some. I'll get into it.

Steve Ives

[tscaff]

Thanks for the consideration Steve. I realize it could be a very big ask. Thanks again.