json - Project subdocument elements when element name not known in Mongo -

i have json file imported mongo:

{   "people": {       "employee1234": {           "salary": 10000,           "dept": "accounting"        },        "employee1235": {            "salary": 40000,            "dept": "ceo"        },        ...   } } 

i want able find of unique salaries , departments of people.

this tricky because people aren't in [] id field, rather elements eid. i'm trying equivalent of find({},{people.*.dept}), can't * wildcard. how can query (given schema of existing documents) ?

that quite impossible mongodb query current schema have dynamic keys. however, suggest change schema such keys become values , store them in embedded document. schema easy query:

{     "people": [         {             "name": "employee1234",             "salary": 10000,             "dept": "accounting"         },             "name": "employee1235",             "salary": 40000,             "dept": "ceo"         }            ] } 

converting current schema 1 above use native javascript:

db.collection.find().foreach(function(doc){     var people = [],         keys = object.keys(doc.people);         obj = {};      keys.foreach(function(key){         obj = doc.people[key];         obj.name = key         people.push(obj);     });      doc.people = people;     db.collection.save(doc); }); 

after changing schema, able query using aggregation framework. aggregation pipeline find of unique salaries , departments of people follows:

db.collection.aggregate([     {         "$unwind": "$people"     },     {         "$group": {             "_id": {                 "salary": "$people.salary",                 "department": "$people.dept"             },             "count": {                 "$sum": 1             }         }     },     {         "$match": { "count": 1 }     } ]); 

for above sample document, output be:

/* 0 */ {     "result" : [          {             "_id" : {                 "salary" : 40000,                 "department" : "ceo"             },             "count" : 1         },          {             "_id" : {                 "salary" : 10000,                 "department" : "accounting"             },             "count" : 1         }     ],     "ok" : 1 }