I have some very large documents. Due to embedados and arrays fields.
An example follows:

{"_id" => BSON :: ObjectId ('5654b4b3204528cfa0000001'),
 "rating" => 1,
 "status" =>: ok,
 "bound_count" => 5,
 "group_ids" =>
  [BSON :: ObjectId ('55ba5d5420452896a8000007'),
   BSON :: ObjectId ('5602bf4c2045289107000001')]
"events": [
{"_id" => BSON :: ObjectId ('5654b4b3204528cfa0000002'),
    "field_a" =>: died,
    "field_b" => {},
    "EVENT_DATE" => 11.24.2015 19:04:19 UTC
    "trigger" => true,
    'type' =>: INLIST,
    "list_id" => BSON :: ObjectId ('5602bf4c2045289107000001'),
    "updated_at" => 11.24.2015 19:04:19 UTC
    "created_at" => 11.24.2015 19:04:19 UTC}
]}

There may be hundreds of events and no more than ten group_ids for example.

For my application I need some fields, but do not need the events.

In version 3 Mongoid I filtered the necessary fields, performing my actions and alter some of these fields returned.

entry = Model.without (: events) .first

1. Logic of my application
   entry.rating = 3
   entry.status = :error
   entry.save
2. raise Mongoid::Errors::ReadonlyDocument

But Mongoid 4: "Documents now loaded from criteria using #only or #without will now raise an error When Attempting to save".

I can not work it out with atomic operations. Because my model has callbacks (after_save, before_update) that depend on previous states of fields (ActiveModel :: Dirty).

They can help me with any suggestions?