I save my transaction with something like :
{code: \"A\", total: 250000, timestamp: ISODate(\"2016-01-20T23:57:05.771Z\")},
{code: \"B\", total: 300000, time
MongoDB 3.6 added timezone parameter to the date manipulation operators. See Kevin's answer.
We can add the "timestamp" to 7 * 60 * 60 * 1000
in a $project stage.
The following pipeline seems to work in MongoDB 3.4 or older.
db.collection.aggregate([
{ "$project": {
"year": { "$year": { "$add": [ "$timestamp", 7 * 60 * 60 * 1000 ] } },
"month": { "$month": { "$add": [ "$timestamp", 7 * 60 * 60 * 1000 ] } },
"day": { "$dayOfMonth": { "$add": [ "$timestamp", 7 * 60 * 60 * 1000 ] } }
} }
])
As an update, MongoDB 3.6 has a new timezone parameter for date manipulation in the aggregation framework. Most date-related operators accept this optional parameter, see $hour for one example.
For example, if we have a document where the date is exactly the new year in UTC:
> db.test.find()
{"_id": 1, "dt": ISODate("2018-01-01T00:00:00Z")}
We can display the date in New York timezone:
> db.test.aggregate([
... {$project:{
... date:{$dayOfMonth:{date:'$dt',timezone:'America/New_York'}},
... month:{$month:{date:'$dt',timezone:'America/New_York'}},
... year:{$year:{date:'$dt',timezone:'America/New_York'}},
... hour:{$hour:{date:'$dt',timezone:'America/New_York'}}
... }}
... ])
{ "_id": 1, "date": 31, "month": 12, "year": 2017, "hour": 19 }
We can also display the date in Sydney timezone:
> db.test.aggregate([
... {$project:{
... date:{$dayOfMonth:{date:'$dt',timezone:'Australia/Sydney'}},
... month:{$month:{date:'$dt',timezone:'Australia/Sydney'}},
... year:{$year:{date:'$dt',timezone:'Australia/Sydney'}},
... hour:{$hour:{date:'$dt',timezone:'Australia/Sydney'}}
... }}
... ])
{ "_id": 1, "date": 1, "month": 1, "year": 2018, "hour": 11 }
The timezone description is using the standard Olson Timezone Identifier string.