Generate view and query for nested dictionary in document

[olcoz8]

Generate view and query for nested dictionary in document

Expected Behavior

Having documents following this schema:

{
  "_id": "xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx",
  "_rev": "x-xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx",

  "field_1": "something",
  "field_2": "something",
  "field_3": "something",

  "field_of_interest": {
    "id-0": {
      "own_id": "id-0",
      "sub_field_of_interest": "data_a",
      "other_sub_field": "something"
    },
    "id-1": {
      "own_id": "id-1",
      "sub_field_of_interest": "data_b",
      "other_sub_field": "something"
    }
  },

  "field_4": something
}

It might be possible to create a view emitting a sub-field of the nested dictionary of the document for getting a query like this:

"data_a": "id-0"
"data_b": "id-1"

Current Behavior

I didn't find any option for working with nested dictionaries. All the references similar to this issue addressed nested lists in the document, not a nested dictionary.

Possible Solution

Is it possible to work with views and queries with this kind of documents?

Your Environment

Currently, I am working with couchdb through the Python library for Cloudant and CouchDB.

[rjharmon]

Agree. Would like to be able to address the values in a dictionary (object) field items with same selector semantics as arrays. If we could address items.$values as array-like, and use $elemMatch/$allMatch on it, that would be great.

Although a map/reduce index can iterate the values and emit() items to an index (then query with startkey/endkey), accessing the same with a JSON Mango index would be tres helpful.

Alternatively: Is there any semantics/metadata that could be applied to such a view-definition, that would give Mango some kind of access to the map index, perhaps with a spoofed fieldname that can be used in Mango query?