python - project nested columns in mongodb -

i have collection in mongo data looks this:

{     "myinput": "myinput",     "myoutput": {         "result": {             "crossdata": {                 "crossdto": [                     {                         "col1": "11",                         "col2": "12",                     },                     {                         "col1": "21",                         "col2": "22",                     },                     {                         "col1": "31",                         "col2": "32",                     }                 ],                 "reqpartnumber": "myinput"             }         },         "status": {             "code": "0",             "message": "successful operation",             "success": "true"         }     } } 

i trying create mapping table in sql data looks crossdto exist if status success:

myinput  11   12 myinput  21   22 myinput  31   32 ... 

i trying come mongodb query on query side as can don't end crazy nested structure in python convert table format.

what have done far:

db.collection.find({'myoutput.status.success':'true'}, {_id:0, myoutput:1}).limit(1) 

and select records status success , meanwhile, show "myoutput"..

but want few levels deep, this:

{     "crossdto": [         {             "col1": "1",             "col2": "1",         },         {             "col1": "2",             "col2": "2",         },         {             "col1": "3",             "col2": "3",         }     ],     "reqpartnumber": "myinput" } 

can me ? how can configure projection part of mongodb can nest columns want... expect similar this:

{_id=0, myoutput.result.crossdata=1} 

thanks :)

the best way in mongodb use aggregation framework. framework pipeline allows query , transform data. in case, you'll want this:

db.collection.aggregate({     $match: {         'myoutput.status.success':'true'     } }, {     $project: {         'crossdto': '$myoutput.result.crossdata.crossdto',         'reqpartnumber': '$myoutput.result.crossdata.reqpartnumber'     } }) 

doing in python isn't bad, , how you'll have if have more data aggregation framework can handle:

mydata = [doc['myoutput']['result']['crossdata'] doc in db.collection.find({'myoutput.status.success':'true'}, {_id:0, myoutput:1})]