Re: [apps-discuss] JSON Patch not suitable for GeoJSON?

On Thu, Jan 24, 2013 at 5:00 AM, Martin Kofahl <M.Kofahl@gmx.de> wrote:

> Hello. I came across the draft-ietf-appsawg-json-patch-10 proposal and
> hope that's the right here place to put a question.
>
> As far I can see any remove, replace, etc. operations can be applied on
> object members or numbered array elements, only. I don't think this is
> sufficient for some formats. Taking GeoJSON
>
> { "type": "FeatureCollection",
>   "features": [
>     { "type": "Feature",
>       "id": "2",
>       "geometry": {"type": "Point", "coordinates": [102.0, 0.5]},
>       "properties": {"prop0": "value0"}
>     },
>     ...
>   ]
> }
>
> as an example, operations may have to be applied on objects filtered by
> some key-value-conditions in its leafs.
>
> A JSON Patch document may have to look like this:
>
> [
>   { "op": "remove", "path": "/features/*/[\"id\"=\"2\"]::parent" }
> ]
>
> Maybe you can exert this suggestion.
>
>
Take a look at
http://tools.ietf.org/html/draft-snell-json-test-05#section-2.5.1

[
  {
    "op": "remove",
    "path": "/features/0",
    "if": {
      "op": "test",
      "path": "/features/0/id",
      "value": "2"
    }
]

Yes, I know it's not quite the same as your example but it's closer. I
could get it closer by defining an iterator operation... e.g.

[
  {
    "op": "each",
    "path": "/features",
    "apply": {
      "op": "remove",
      "if": {
        "op": "test",
        "path": "/id",
        "value": "2"
      }
    }
  }
]

The way to interpret this would be: For each member of /features (whether
it's an array or object), set the member as the context object and apply
the remove operation only if the condition holds for that item.

Alternatively, we can use the json-predicates draft as the base to
introduce a more complete pointer expression syntax... e.g.

[
  { "op": "remove", "exp": "/features/*/[\"id\"=\"2\"]::parent" }
]

Note the use of "exp" instead of "path"... the hope with json-pointer is to
coexist with json-patch, not replace it...

- James

> Kind regards,
> Martin
> _______________________________________________
> apps-discuss mailing list
> apps-discuss@ietf.org
> https://www.ietf.org/mailman/listinfo/apps-discuss
>