How to group data by every hour - mongodb

Question

How do I get counts data grouped by every hour in 24 hours even if data is not present i.e. IF 0 will select 0
MonogDB 3.6
Input
[
{
"_id": ObjectId("5ccbb96706d1d47a4b2ced4b"),
"date": "2019-05-03T10:39:53.108Z",
"id": 166,
"update_at": "2019-05-03T02:45:36.208Z",
"type": "image"
},
{
"_id": ObjectId("5ccbb96706d1d47a4b2ced4c"),
"date": "2019-05-03T10:39:53.133Z",
"id": 166,
"update_at": "2019-05-03T02:45:36.208Z",
"type": "image"
},
{
"_id": ObjectId("5ccbb96706d1d47a4b2ced4d"),
"date": "2019-05-03T10:39:53.180Z",
"id": 166,
"update_at": "2019-05-03T20:45:36.208Z",
"type": "image"
},
{
"_id": ObjectId("5ccbb96706d1d47a4b2ced7a"),
"date": "2019-05-10T10:39:53.218Z",
"id": 166,
"update_at": "2019-12-04T10:45:36.208Z",
"type": "image"
},
{
"_id": ObjectId("5ccbb96706d1d47a4b2ced7b"),
"date": "2019-05-03T10:39:53.108Z",
"id": 166,
"update_at": "2019-05-05T10:45:36.208Z",
"type": "image"
},
{
"_id": ObjectId("5ccbb96706d1d47a4b2cedae"),
"date": "2019-05-03T10:39:53.133Z",
"id": 166,
"update_at": "2019-05-05T10:45:36.208Z",
"type": "image"
},
{
"_id": ObjectId("5ccbb96706d1d47a4b2cedad"),
"date": "2019-05-03T10:39:53.180Z",
"id": 166,
"update_at": "2019-05-06T10:45:36.208Z",
"type": "image"
},
{
"_id": ObjectId("5ccbb96706d1d47a4b2cedab"),
"date": "2019-05-10T10:39:53.218Z",
"id": 166,
"update_at": "2019-12-06T10:45:36.208Z",
"type": "image"
}
]
Implementation
db.collection.aggregate({
$match: {
update_at: {
"$gte": "2019-05-03T00:00:00.0Z",
"$lt": "2019-05-05T00:00:00.0Z"
},
id: {
"$in": [
166
]
}
}
},
{
$group: {
_id: {
$substr: [
"$update_at",
11,
2
]
},
count: {
"$sum": 1
}
},
},
{
$project: {
_id: 0,
hour: "$_id",
count: "$count"
}
},
{
$sort: {
hour: 1
}
})
Actual Output:
{
"count": 2,
"hour": "02"
},
{
"count": 1,
"hour": "20"
}
My expectation code show 24 hours event data is 0 or null and convert from example "02" as "02 AM" , "13" as "01 PM":
Expected Output
{
"count": 0,
"hour": "01" // 01 AM
},
{
"count": 2,
"hour": "02"
},
{
"count": 0,
"hour": "03"
},
{
"count": 0,
"hour": "04"
},
{
"count": 0,
"hour": "05"
},
{
"count": 1,
"hour": "20" // to 08 pm
}

Answer 1

Try this solution:
Explanation
We group by hour to count how many images are uploaded.
Then, we add extra field hour to create time interval (if you had v4.x, there is a better solution).
We flattern hour field (will create new documents) and split first 2 digits to match count and split last 2 digits to put AM / PM periods.
db.collection.aggregate([
{
$match: {
update_at: {
"$gte": "2019-05-03T00:00:00.0Z",
"$lt": "2019-05-05T00:00:00.0Z"
},
id: {
"$in": [
166
]
}
}
},
{
$group: {
_id: {
$substr: [
"$update_at",
11,
2
]
},
count: {
"$sum": 1
}
}
},
{
$addFields: {
hour: [
"0000",
"0101",
"0202",
"0303",
"0404",
"0505",
"0606",
"0707",
"0808",
"0909",
"1010",
"1111",
"1212",
"1301",
"1402",
"1503",
"1604",
"1705",
"1806",
"1907",
"2008",
"2109",
"2210",
"2311"
]
}
},
{
$unwind: "$hour"
},
{
$project: {
_id: 0,
hour: 1,
count: {
$cond: [
{
$eq: [
{
$substr: [
"$hour",
0,
2
]
},
"$_id"
]
},
"$count",
0
]
}
}
},
{
$group: {
_id: "$hour",
count: {
"$sum": "$count"
}
}
},
{
$sort: {
_id: 1
}
},
{
$project: {
_id: 0,
hour: {
$concat: [
{
$substr: [
"$_id",
2,
2
]
},
{
$cond: [
{
$gt: [
{
$substr: [
"$_id",
0,
2
]
},
"12"
]
},
" PM",
" AM"
]
}
]
},
count: "$count"
}
}
])
MongoPlayground

Answer 2

There's no "magic" solution, you'll have to hardcode it into your aggregation:
Heres an example using Mongo v3.2+ syntax with some $map and $filter magic:
db.collection.aggregate([
{
$match: {
update_at: {
"$gte": "2019-05-03T00:00:00.0Z",
"$lt": "2019-05-05T00:00:00.0Z"
},
id: {"$in": [166]}
}
},
{
$group: {
_id: {$substr: ["$update_at", 11, 2]},
count: {"$sum": 1}
}
},
{
$group: {
_id: null,
hours: {$push: {hour: "$_id", count: "$count"}}
}
},
{
$addFields: {
hours: {
$map: {
input: {
$concatArrays: [
"$hours",
{
$map: {
input: {
$filter: {
input: ["00", "01", "02", "03", "04", "05", "06", "07", "08", "09", "10", "11", "12", "13", "14", "15", "16", "17", "18", "19", "20", "21", "22", "23"],
as: "missingHour",
cond: {
$not: {
$in: [
"$$missingHour",
{
$map: {
input: "$hours",
as: "hourObj",
in: "$$hourObj.hour"
}
}
]
}
}
}
},
as: "missingHour",
in: {hour: "$$missingHour", count: 0}
}
}
]
},
as: "hourObject",
in: {
count: "$$hourObject.count",
hour: {
$cond: [
{$eq: [{$substr: ["$$hourObject.hour", 0, 1]}, "0"]},
{$concat: ["$$hourObject.hour", " AM"]},
{
$concat: [{
$switch: {
branches: [
{case: {$eq: ["$$hourObject.hour", "13"]}, then: "1"},
{case: {$eq: ["$$hourObject.hour", "14"]}, then: "2"},
{case: {$eq: ["$$hourObject.hour", "15"]}, then: "3"},
{case: {$eq: ["$$hourObject.hour", "16"]}, then: "4"},
{case: {$eq: ["$$hourObject.hour", "17"]}, then: "5"},
{case: {$eq: ["$$hourObject.hour", "18"]}, then: "6"},
{case: {$eq: ["$$hourObject.hour", "19"]}, then: "7"},
{case: {$eq: ["$$hourObject.hour", "20"]}, then: "8"},
{case: {$eq: ["$$hourObject.hour", "21"]}, then: "9"},
{case: {$eq: ["$$hourObject.hour", "22"]}, then: "10"},
{case: {$eq: ["$$hourObject.hour", "23"]}, then: "11"},
],
default: "None"
}
}, " PM"]
}
]
}
}
}
}
}
},
{
$unwind: "$hours"
},
{
$project: {
_id: 0,
hour: "$hours.hour",
count: "$hours.count"
}
},
{
$sort: {
hour: 1
}
}
]);
A short explanation of the $addFields stage: we first add hours that we're missing, we then merge the two arrays (of the original found hours and the "new" missing hours), finally we convert to the required output ("01" to "01 AM").
If you're using Mongo v4+ I recommend you change the $group _id stage to use $dateFromString as its more consistent.
_id: {$hour: {$dateFromString: {dateString: "$update_at"}}}
If you do do that, you'll have to update the $filter and $map section to use numbers and not strings and eventually using $toString to cast into the format you want, hence the v4+ requirement.

Answer 3

You should store date values as Date objects instead of strings. I would do the formatting like this:
db.collection.aggregate(
[
{ $match: { ... } },
{
$group: {
_id: { h: { $hour: "$update_at" } },
count: { $sum: 1 }
}
},
{
$project: {
_id: 0,
hour: {
$switch: {
branches: [
{ case: { $lt: ["$_id.h", 10] }, then: { $concat: ["0", { $toString: "$_id.h" }, " AM"] } },
{ case: { $lt: ["$_id.h", 13] }, then: { $concat: [{ $toString: "$_id.h" }, " AM"] } },
{ case: { $lt: ["$_id.h", 22] }, then: { $concat: ["0", { $toString: { $subtract: ["$_id.h", 12] } }, " PM"] } },
{ case: { $lt: ["$_id.h", 24] }, then: { $concat: [{ $toString: { $subtract: ["$_id.h", 12] } }, " PM"] } }
]
}
},
hour24: "$_id.h",
count: 1
}
},
{ $sort: { hour24: 1 } }
])
As non-American I am not familiar with AM/PM rules, esp. for midnight and midday but I guess you get the principle.

Answer 4

Here is the query you can test it out, for MongoDB 4.0+
i will be improving query and update
const query = [{
$match: {
update_at: {
"$gte": ISODate("2019-05-03T00:00:00.0Z"),
"$lt": ISODate("2019-05-05T00:00:00.0Z")
},
id: {
"$in": [
166
]
}
}
},
{
$group: {
_id: { $hour: "$update_at" },
count: {
"$sum": 1
}
},
},
{
$addFields: {
hourStr: { $toString: { $cond: { if: { $gte: ["$_id", 12] }, then: { $subtract: [12, { $mod: [24, '$_id'] }] }, else: "$_id" } } },
}
},
{
$project: {
formated: { $concat: ["$hourStr", { $cond: { if: { $gt: ["$_id", 12] }, then: " PM", else: " AM" } }] },
count: "$count",
hour: 1,
}
}]

Answer 5

If you want to output in Indian Time formate. then below code work!
const query = [
{
$match: {
update_at: {
"$gte": ISODate("2019-05-03T00:00:00.0Z"),
"$lt": ISODate("2019-05-05T00:00:00.0Z")
},
id: {
"$in": [
166
]
}
}
},
{
$project: {
"h": { "$hour": { date: "$update_at", timezone: "+0530" } },
}
},
{
$group:
{
_id: { $hour: "$h" },
count: { $sum: 1 }
}
}
];