If all you need is serializing mongo results into json, it is possible to use the json
module, provided you define custom handlers to deal with non-serializable fields types.
One advantage is that you have full power on how you encode specific fields, like the datetime representation.
Here is a handler which encodes dates using the iso representation and the id as an hexadecimal string:
import pymongo
import json
import datetime
import bson.objectid
def my_handler(x):
if isinstance(x, datetime.datetime):
return x.isoformat()
elif isinstance(x, bson.objectid.ObjectId):
return str(x)
else:
raise TypeError(x)
db = pymongo.MongoClient().samples
record = db.movies.find_one()
# {u'_id': ObjectId('5692a15524de1e0ce2dfcfa3'), u'title': u'Toy Story 4',
# u'released': datetime.datetime(2010, 6, 18, 4, 0),}
json_record = json.dumps(record, default=my_handler)
# '{"_id": "5692a15524de1e0ce2dfcfa3", "title": "Toy Story 4",
# "released": "2010-06-18T04:00:00"}'