oneOf in schema.json is showing null in generated documentaion

[sammyApe]

No description provided.

[geemus]

You could probably do this via custom templates for the docs, but I'm not sure I totally know what you want. Could you provide a small-ish sample schema that demonstrates what you mean?

[sammyApe]

so with a schema.json like the one below, If I execute the doc command i get null for a but b comes as expected

{
  "$schema": "http://interagent.github.io/interagent-hyper-schema",
  "type": [
    "object"
  ],
  "definitions": {
    "a": {
      "title": "a",
      "description": "",
      "type": [
        "object"
      ],
      "oneOf": [
        {
          "properties": {
            "c": {
              "type":"object",
              "description":"c element"
            }
          },
          "additionalProperties": false
        },
        {
          "properties": {
            "d": {
              "type":"object",
              "description":"d element"

            }
          },
          "additionalProperties": false
        }
      ],
      "definitions":{
        "identity": {
        }
      },
      "links":[]

    },
    "b":{
      "title": "a",
      "type":"string",
      "description": "b element",
      "definitions":{
        "identity": {
        }
      },
      "links":[]
    }

  },
  "properties": {
    "a": {
      "$ref": "#/definitions/a"
    },
    "b": {
      "$ref": "#/definitions/b"
    }
  },
  "links": [
    {
      "href": "https://xxx.com",
      "rel": "self"
    }
  ],
  "id": "xxx",
  "description":"xxx element",
  "title": "",
  "additionalProperties": false
}

[geemus]

Ah. I wonder if you might have better luck if you reversed the nesting. So a would have the oneOf inside the properties attribute (instead of having the properties attributes repeated inside). Worth a shot anyway.

[sammyApe]

I tried that way but it only works when c and d enums .. I had to tweak the schemata.md.erb file to make it work for me. Thanks

[geemus]

@sammyApe thanks for the update, glad to hear you found a workaround. Hopefully we can get this fixed/improved in the future so that wouldn't be needed.