mongodb - Inline/combine other collection into one collection -

i want combine 2 mongodb collections.

basically have collection containing documents reference 1 document collection. want have inline / nested field instead of separate document.

so provide example:

collection a:

[{     "_id":"90a26c2a-4976-4edd-850d-2ed8bea46f9e",     "somevalue": "foo"   },   {     "_id":"5f0bb248-e628-4b8f-a2f6-fecd79b78354",     "somevalue": "bar"   }] 

collection b:

[{     "_id":"169099a4-5eb9-4d55-8118-53d30b8a2e1a",     "collectionaid":"90a26c2a-4976-4edd-850d-2ed8bea46f9e",     "some":"foo",     "andother":"stuff"   },   {     "_id":"83b14a8b-86a8-49ff-8394-0a7f9e709c13",     "collectionaid":"90a26c2a-4976-4edd-850d-2ed8bea46f9e",     "some":"bar",     "andother":"random"    }] 

this should result in collection looking this:

[{     "_id":"90a26c2a-4976-4edd-850d-2ed8bea46f9e",     "somevalue": "foo",     "collectionb":[{             "some":"foo",             "andother":"stuff"             },{             "some":"bar",             "andother":"random"             }]   },   {     "_id":"5f0bb248-e628-4b8f-a2f6-fecd79b78354",     "somevalue": "bar"   }] 

i'd suggest simple console:

db.collb.find().foreach(function(doc) {     var aid = doc.collectionaid;     if (typeof aid === 'undefined') { return; } // nothing     delete doc["_id"]; // remove property     delete doc["collectionaid"]; // remove property     db.colla.update({_id: aid},   /* match id b */        { $push : { collectionb : doc }}); }); 

it loops through each document in collectionb , if there field collectionaid defined, removes unnecessary properties (_id , collectionaid). finally, updates matching document in collectiona using $push operator add document b field collectionb. if field doesn't exist, automatically created array newly inserted document. if exist array, appended. (if exists, isn't array, fail). because update call isn't using upsert, if _id in collectionb document doesn't exist, nothing happen.

you can extend delete other fields necessary or possibly add more robust error handling if example document b doesn't match in a.

running code above on data produces this:

{ "_id" : "5f0bb248-e628-4b8f-a2f6-fecd79b78354", "somevalue" : "bar" } { "_id" : "90a26c2a-4976-4edd-850d-2ed8bea46f9e",     "collectionb" : [             {                     "some" : "foo",                     "andother" : "stuff"             },             {                     "some" : "bar",                     "andother" : "random"             }     ],     "somevalue" : "foo" }