Pymongo將時間戳轉換為新字段

Question

我的集合中有很多實體，我必須在集合中創建一個新的日期字段以用於將來的查詢。

{'_id': ObjectId('5afea920d326051990a7f337'), 'created_at': 'Fri May 18 10:21:07 +0000 2018', 'timestamp_ms': '1526638867739'}
{'_id': ObjectId('5afea920d326051990a7f339'), 'created_at': 'Fri May 18 10:21:08 +0000 2018', 'timestamp_ms': '1526638868310'}
{'_id': ObjectId('5afea972d326051c5c05bc11'), 'created_at': 'Fri May 18 10:22:30 +0000 2018', 'timestamp_ms': '1526638950799'}
{'_id': ObjectId('5afea974d326051c5c05bc16'), 'created_at': 'Fri May 18 10:22:32 +0000 2018', 'timestamp_ms': '1526638952160'}
{'_id': ObjectId('5afea974d326051c5c05bc17'), 'created_at': 'Fri May 18 10:22:32 +0000 2018', 'timestamp_ms': '1526638952841'}

我需要將timestamp_ms轉換為日期格式，如下所示：

{'_id': ObjectId('5afea920d326051990a7f337'), 'created_at': 'Fri May 18 10:21:07 +0000 2018', 'timestamp_ms': '1526638867739’, 'NewDate': '2018 05 18 10:21:07'}
{'_id': ObjectId('5afea920d326051990a7f339'), 'created_at': 'Fri May 18 10:21:08 +0000 2018', 'timestamp_ms': '1526638868310’, 'NewDate': '2018 05 18 10:21:08'}
{'_id': ObjectId('5afea972d326051c5c05bc11'), 'created_at': 'Fri May 18 10:22:30 +0000 2018', 'timestamp_ms': '1526638950799’, 'NewDate': '2018 05 18 10:22:30'}
{'_id': ObjectId('5afea974d326051c5c05bc16'), 'created_at': 'Fri May 18 10:22:32 +0000 2018', 'timestamp_ms': '1526638952160’, 'NewDate': '2018 05 18 10:22:32'}
{'_id': ObjectId('5afea974d326051c5c05bc17'), 'created_at': 'Fri May 18 10:22:32 +0000 2018', 'timestamp_ms': '1526638952841’, 'NewDate': '2018 05 18 10:22:32'}

我使用了這段代碼（Python 3.6，pymongo 3.8，mongodb 4.0）：

pipeline = [
    {
        '$addFields': {
            'newDate': {
                '$toDate': '$timestamp_ms'
            }
        }
    }
]
cursor = collection.aggregate(pipeline)

但是給出了以下錯誤消息： pymongo.errors.OperationFailure: Error parsing date string '1526638867739'; 12: Unexpected character '9' pymongo.errors.OperationFailure: Error parsing date string '1526638867739'; 12: Unexpected character '9'

我不確定聚合是否是執行此任務的正確方法。 datetime.strptime()對於'created_at'可能更好，但我還沒想出如何將它實現到db.Mycollection_update_many() 。

Answer 1

使用下面的查詢與pymongo和mongodb 4.0

db.test.aggregate(
    [
        {
            "$addFields": {
                "NewDate": {"$toDate": "$timestamp_ms"} 
            }
        },
        {
            "$out": "test"
        },
    ],
)

Answer 2

我從MongoDB大學討論論壇的Kanika Singla那里得到了我的問題答案。 答案就在這里，如果你有同樣的問題。

pipeline = [
    {
        '$project': {
            'yearMonthDayUTC': {
                '$convert': {
                    'to': 'double',
                    'input': '$timestamp_ms'
                }
            }
        }
    }, {
        '$addFields': {
            'newDate': {
                '$toDate': '$yearMonthDayUTC'
            }
        }
    }
]