Data type conversion in MongoDB

Question

I have a collection called Document in MongoDB. Documents in this collection have a field called CreationDate stored in ISO date type. My task is to count the number of docu

Answer 1

To convert date format to "xxxx-xx-xx"

db.getCollection('analytics').aggregate([
 {$project: {
     day: {$dayOfMonth: "$createdAt"},
     month: {$month: "$createdAt"},
     year: {$year: "$createdAt"},
     }
  },
  {$project: {
      day: {$concat:["0", {$substr:["$day",0, 2]}]},
     month: {$concat:["0", {$substr:["$month",0, 2]}]},
     year: {$substr:["$year",0, 4]},
      }
  },
  {$project: {date: {$concat: [{$substr:["$year",0, 4]},"-", {$substr:["$month",0, 2]},"-", {$substr:["$day",0, 2]}]}}},
]);

Answer 2

FOR MONGO >= 3.0 (after ~2015)

In case some poor souls stumble across this question in 2017, like me: As of Mongo 3.0 there is now a dateToString operator available

which means if you have proper Date()'s you should be able to do simply:

db.Document.aggregate( 
    , {$project: {_id:1, CreationDate:1}
    , {$group: {
         _id : { $dateToString: { format: "%Y-%m-$d", date: "$CreationDate" } },
         cnt:{$sum:1}}}
    , {$sort:{'cnt':-1}}
);

For those of us who happen to have dates stored in non-date field (what joy!) you can create a new Date() object in the project step. In my case, the date was stored as a number of milliseconds (integer) and I add the number of milliseconds to the 0-Date.

db.Document.aggregate( 
    , {$project: {
         _id:1, 
         CreationDate: {"$add": [ new Date(0), "$CreatedOn" ]}
}
    , {$group: {
         _id : { $dateToString: { format: "%Y-%m-$d", date: "$CreationDate" } },
         cnt:{$sum:1}}}
    , {$sort:{'cnt':-1}}
);

Answer 3

You can do this with $concat but first you need to convert to a string via $substr, also handling the double digit case:

db.Document.aggregate([ 
    { "$group": { 
        "_id":{ 
            "$concat": [
                 { "$substr": [ { "$year": "$CreationDate" }, 0, 4 ] },
                 "-",
                 { "$cond": [
                     { "$gt": [ { "$month": "$CreationDate" }, 9 ] },
                     { "$substr": [ { "$month": "$CreationDate" }, 0, 2 ] },
                     { "$concat": [
                         "0",
                         { "$substr": [ { "$month": "$CreationDate" }, 0, 1 ] },
                     ]},
                 ]},
                 "-",
                 { "$cond": [
                     { "$gt": [ { "$dayOfMonth": "$CreationDate" }, 9 ] },
                     { "$substr": [ { "$dayOfMonth": "$CreationDate" }, 0, 2 ] },
                     { "$concat": [
                         "0",
                         { "$substr": [ { "$dayOfMonth": "$CreationDate" }, 0, 1 ] },
                     ]}
                 ]}
             ]
         },
         { "cnt": { "$sum": 1 } }
    }}
    { "$sort":{ "cnt" :-1 }}
]);

Possibly better is to just use date math instead, this returns an epoch timestamp value, but it is easy to work into a date object in post processing:

db.Document.aggregate([
    { "$group": {
        "_id": {
            "$subtract": [
                { "$subtract": [ "$CreationDate", new Date("1970-01-01") ] },
                { "$mod": [
                    { "$subtract": [ "$CreationDate", new Date("1970-01-01") ] },
                    1000 * 60 * 60 * 24
                ]}
            ]
        },
        "cnt": { "$sum": 1 }
    }},
    { "$sort": { "cnt": -1 } }
])

Answer 4

Another easier way you can convert ISODate to various date formats is to use the $dateToString aggregation operator.

db.collection.aggregate([
    { $group: {
        _id: { $dateToString: { format: "%Y-%m-%d %H:%M", date: "$CreationDate" } },
        count: { $sum: 1 }
    }},
    { $sort : { count: -1 }}
])

Answer 5

Neil's answer its right to the point but has a little bug. In the conditionals (to check if the number of the month and day needs to be prepended with a 0) is not $gt but $lte.

With $lte, only months and days with one digit will be prepended with 0.

e.g: 2014-10-03, 2014-08-03.