JSON array names about edi-json HOT 3 OPEN

Is your suggestion to avoid arrays completely in the JSON? There are a number of things in EDI that are naturally repeated, without any visible identifier for each repetition. For example, transactions within a functional group, or functional groups within an interchange. Or repetitions of a segment loop. Representing these as JSON arrays seems to me the obvious and simplest choice. Using your example, if there were multiple N1-1000A loops (I realize that this particular loop is not repeated but some loops are) how would that work? the keys need to be unique.

from edi-json.

I'm attempting to create a schema for this json format for a specific X12 transaction type and version, and running into challenges that I think are directly related to your point about the anonymous arrays. I am experimenting with specific alternative formats, looking to choose one that is is more convenient for the type of navigation you have in mind, well-suited for transaction-specific schemas, and yet is reasonably readable by humans.

Would you be willing to give feedback/comments on the various alternatives I am considering? I am envisioning a major new release that continues to support the original "light" format as well as a new format; it would be important to get some trial use first. If you are interested, please contact me separately. Thanks.

from edi-json.

Yes! In the ideal state, we should be able to take a json document and convert it into a javascript object model that is simple to navigate. I could write you a test case that demonstrates it and let you provide the implementation that makes it work.

from edi-json.