I am trying to
get all the ids for which the period.startDate > sysdate
get all the ids for which the period.endDate < sysdate
from the JSON.
[
{
"id": 1,
"period":
[
{
"startDate": "2020-05-05",
"endDate": "2020-05-06"
},
{
"startDate": "2020-06-05",
"endDate": "2020-06-06"
}
]
},
{
"id": 2,
"period":
[
{
"startDate": "2024-07-05",
"endDate": "2024-07-06"
},
{
"startDate": "2024-08-05",
"endDate": "2024-08-06"
}
]
}
]
I have tried to go far as below aggregation:
[{
$project: {
_id: 0,
sId: '$id',
period: 1
} }, {
$unwind: {
path: '$period',
includeArrayIndex: 'index'
} }, {
$group: {
_id: '$sId',
minDate: {
$min: '$periods.startDate'
}
} }, {
$project: {
storeId: '$_id',
_id: 0,
minDated: {
$dateFromString: {
dateString: '$minDate'
}
},
today: ISODate('2022-08-03T11:37:03.954Z')
} }]
One option is using $reduce and $group:
db.collection.aggregate([
{$project: {
_id: 0,
id: 1,
minDate: {
$dateFromString: {
dateString: {
$reduce: {
input: "$period",
initialValue: {$first: "$period.startDate"},
in: {$min: ["$$value", "$$this.startDate"]}
}
}
}
},
maxDate: {
$dateFromString: {
dateString: {
$reduce: {
input: "$period",
initialValue: {$first: "$period.endDate"},
in: {$max: ["$$value", "$$this.startDate"]}
}
}
}
}
}
},
{$group: {
_id: 0,
startDateLargerIds: {
$push: {
$cond: [{$gt: ["$minDate", ISODate("2022-08-03T11:37:03.954Z")]},
"$id", "$$REMOVE"]}
},
endDateSmallerIds: {
$push: {
$cond: [{$lt: ["$maxDate", ISODate("2022-08-03T11:37:03.954Z")]},
"$id", "$$REMOVE"]}
}
}
},
{$unset: "_id"}
])
See how it works on the playground example
I have a similar collection where I have sort them by their startTime:
{"name": 'A', "startTime": '1634626355', "endTime": '1634631405'}
{"name": 'A', "startTime": '1634631406', "endTime": '1634631864'}
{"name": 'A', "startTime": '1634631865', "endTime": '1634656048'}
{"name": 'A', "startTime": '1634712642', "endTime": '1634718856'}
How can I compare the documents such that if the document endTime and the next document startTime duration is less than 5 minutes, merge it.
This is the result I'm trying to achieve (The 1st 3 documents are merged into 1 where it uses the startTime of the 1st document and the endTime of the 3rd document):
{"name": 'A', "startTime": '1634626355', "endTime": '1634656048'}
{"name": 'A', "startTime": '1634712642', "endTime": '1634718856'}
Thanks
First of all, you should never store date/time values as string, it's a design flaw. Store always proper Date object.
This solution works without self-lookup, so it may perform better:
db.collection.aggregate([
{
$set: {
startDateTime: { $toDate: { $multiply: ["$startTime", 1000] } },
endDateTime: { $toDate: { $multiply: ["$endTime", 1000] } }
},
},
{ $sort: { startDateTime: 1 } },
{ $group: { _id: null, data: { $push: "$$ROOT" } } },
{
$set: {
data: {
$reduce: {
input: "$data",
initialValue: [],
in: {
$cond: {
if: {
$or: [
{ $eq: [{ $size: "$$value" }, 0] }, // for the initail element
{
$gt: [
{
$dateDiff: { // calculate difference
endDate: "$$this.startDateTime",
startDate: { $last: "$$value.endDateTime" },
unit: "minute"
}
},
5 // more than 5 Minutes
]
}
]
},
then: { $concatArrays: ["$$value", ["$$this"]] }, // append new element
else: {
$map: {
input: "$$value",
as: "data",
in: {
$cond: {
if: { $eq: ["$$data._id", { $last: "$$value._id" }] }, // find last element
then: { // update last element
$mergeObjects: [
"$$data",
{ endDateTime: "$$this.endDateTime" },
{ endTime: "$$this.endTime" }
]
},
else: "$$data"
}
}
}
}
}
}
}
}
}
},
// some cosmetic
{ $unwind: "$data" },
{ $replaceRoot: { newRoot: "$data" } }
])
Mongo Playground
You can use $lookup in an aggregation pipeline to find out the documents that you need to remove. Then, perform a forEach to remove them.
db.collection.aggregate([
{
$addFields: {
endDateTime: {
"$toDate": {
"$multiply": [
{
$toLong: "$endTime"
},
1000
]
}
}
},
},
{
"$lookup": {
"from": "collection",
let: {
end: "$endDateTime"
},
pipeline: [
{
"$addFields": {
startDateTime: {
"$toDate": {
"$multiply": [
{
$toLong: "$startTime"
},
1000
]
}
}
}
},
{
$match: {
$expr: {
$and: [
{
$lte: [
{
$subtract: [
"$startDateTime",
"$$end"
]
},
300000
]
},
{
$lte: [
"$$end",
"$startDateTime"
]
}
]
}
}
}
],
"as": "lessThan5min"
}
},
{
"$unwind": "$lessThan5min"
},
{
"$replaceRoot": {
"newRoot": "$lessThan5min"
}
}
]).forEach(function(doc){
db.collection.remove({ "_id": doc._id });
});
Here is the Mongo playground to find out the documents that you need to remove for your reference.
I have a pipeline that gives me the result according to the players belonging to a certain company in a daily, weekly, and monthly manner. I have a date filter in the dashboard which gives an iso date range. I want to limit or range my results according to the date filter? is there any way to do it in the pipeline?
if (chartType === 'Daily') {
statsPipelineCondition = {
_id: { day: { $dayOfMonth: '$createdAt' }, month: { $month: '$createdAt' }, year: { $year: '$createdAt' } },
};
sortCondition = { '_id.year': 1, '_id.month': 1, '_id.day': 1 };
} else if (chartType === 'Monthly') {
statsPipelineCondition = {
_id: { month: { $month: '$createdAt' }, year: { $year: '$createdAt' } },
};
sortCondition = { '_id.year': 1, '_id.month': 1 };
} else {
statsPipelineCondition = {
_id: { week: { $week: '$createdAt' }, year: { $year: '$createdAt' } },
};
sortCondition = { '_id.year': 1, '_id.week': 1 };
}
const statsPipeline = [
{
$group: {
...statsPipelineCondition,
TOTAL: { $sum: 1 },
XR: { $sum: { $cond: [{ $in: ['$company', ['XR', 'CR', 'DX']] }, 1, 0] } },
CT: { $sum: { $cond: [{ $eq: ['$company', 'CT'] }, 1, 0] } },
MR: { $sum: { $cond: [{ $eq: ['$company', 'MR'] }, 1, 0] } },
MG: { $sum: { $cond: [{ $in: ['$company', ['NM', 'MM', 'MG']] }, 1, 0] } },
},
},
{
$sort: {
...sortCondition,
},
},
];
Date filter:
datefilter - { '$gte': '2020-09-01T04:49:50.899Z',
'$lte': '2020-11-03T04:49:50.899Z' }
You need to add a $match stage prior to the $group stage to filter our based on the range:
let datefilter = {
'$gte': new Date('2020-09-01T04:49:50.899Z'),
'$lte': new Date('2020-11-03T04:49:50.899Z')
};
const statsPipeline = [
{
$match: {
createdAt: datefilter
}
},
{
$group: {
...statsPipelineCondition,
TOTAL: { $sum: 1 },
XR: { $sum: { $cond: [{ $in: ['$company', ['XR', 'CR', 'DX']] }, 1, 0] } },
CT: { $sum: { $cond: [{ $eq: ['$company', 'CT'] }, 1, 0] } },
MR: { $sum: { $cond: [{ $eq: ['$company', 'MR'] }, 1, 0] } },
MG: { $sum: { $cond: [{ $in: ['$company', ['NM', 'MM', 'MG']] }, 1, 0] } },
},
},
{
$sort: {
...sortCondition,
},
},
];
Okay, So for some reason. MongoDB doesn't like dates in strings. It would be much better if we just convert the strings by an operator provided by the aggregation framework and this will make the things work.
{
$match: {
$expr: {
$and: [
{
$gte: [
'$createdAt',
{
$dateFromString: {
dateString: dateFilter.$gte,
},
},
],
},
{
$lte: [
'$createdAt',
{
$dateFromString: {
dateString: dateFilter.$lte,
},
},
],
},
],
},
},
},
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
}
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
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.
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.
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,
}
}]
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: { $sum: 1 }
}
}
];
Dear attempts to filter by date in mongo but I have not succeeded. I pass my query and structure of JSON.
db.getCollection('articles').aggregate([
{ "$match": {
"$expr": {
"$and": [
{
"$gte": [
{ "$dateFromString": { "creationDate": "10-08-2018", "format": "%m-%d-%Y" }}
]
},
{
"$lte": [
{ "$dateFromString": { "creationDate": "10-08-2018", "format": "%m-%d-%Y" }}
]
}
]
}
}}
])
My JSON is
{
"_id" : ObjectId("5bbb6b1de75b933850a608fc"),
"title" : "05",
"body" : "asgfasgasfa",
"creationDate" : ISODate("2018-10-08T14:35:07.000Z"),
"operationType" : "C",
"__v" : 0
}
MongoDB : v3.6.3
If you a looking for a match on "10-08-2018" exactly from a date field you can use $dateToString in combination with $eq:
db.getCollection('articles').aggregate([
{
"$match": {
"$expr": {
$eq: [
"10-08-2018",
{
"$dateToString": {
"date": "$creationDate",
"format": "%m-%d-%Y"
}
}
]
}
}
}
])
See it working here
If you are looking for a set of records matching a date range:
db.getCollection('articles').aggregate([
{
"$match": {
"$expr": {
"$and": [
{
"$gte": [
"$creationDate",
{
"$dateFromString": {
"dateString": "10-08-2018",
"format": "%m-%d-%Y"
}
}
]
},
{
"$lte": [
"$creationDate",
{
"$dateFromString": {
"dateString": "11-08-2018",
"format": "%m-%d-%Y"
}
}
]
}
]
}
}
}
])
See it working here
One note with the 2nd example is that it would do the dates as ISO date so it would not cover the end of day 11-08-2018 but 2018-11-08T00:00:00.000Z as pointed out by matthPen comment.
Even if #Akrion answer is returning sample, it not properly filter dates. In a global way, it's better to deal with dates as often as you can, instead of transforming to string. A simple example : using your format ("%m-%d-%Y"), Date("10-08-2018") > Date("12-01-2017"), but in term of string comparison, "10-08-2018" < "12-01-2017" . The only working format in this case is %Y-%m-%d .
Anyway, mongoDB provides some operators in aggregation framework, to properly work with dates. Here's a query to match by day :
db.collection.aggregate([
{
$match: {
$expr: {
$and: [
{
$eq: [
{
$year: new Date("2018-10-08")
},
{
$year: "$creationDate"
}
]
},
{
$eq: [
{
$month: new Date("2018-10-08")
},
{
$month: "$creationDate"
}
]
},
{
$eq: [
{
$dayOfMonth: new Date("2018-10-08")
},
{
$dayOfMonth: "$creationDate"
}
]
}
]
}
}
}
])
You can try it here
It can be a little more tricky to adapt this query for dates range, but the following query will do the job. You can test it here.
db.collection.aggregate([
{
$match: {
$expr: {
$and: [
{
// Start date definition, included ($gte)
$or: [
{
$and: [
{
$eq: [
{
$year: new Date("2018-10-08")
},
{
$year: "$creationDate"
}
]
},
{
$eq: [
{
$month: new Date("2018-10-08")
},
{
$month: "$creationDate"
}
]
},
{
$gte: [
{
$dayOfMonth: "$creationDate"
},
{
$dayOfMonth: new Date("2018-10-08")
}
]
}
]
},
{
$and: [
{
$eq: [
{
$year: "$creationDate"
},
{
$year: new Date("2018-10-08")
}
]
},
{
$gte: [
{
$month: "$creationDate"
},
{
$month: new Date("2018-10-08")
}
]
},
]
},
{
$and: [
{
$gte: [
{
$year: "$creationDate"
},
{
$year: new Date("2018-10-08")
}
]
},
]
}
],
},
//end date definition, excluded ($lt)
{
$or: [
{
$and: [
{
$eq: [
{
$year: new Date("2018-11-08")
},
{
$year: "$creationDate"
}
]
},
{
$eq: [
{
$month: new Date("2018-11-08")
},
{
$month: "$creationDate"
}
]
},
{
$lt: [
{
$dayOfMonth: "$creationDate"
},
{
$dayOfMonth: new Date("2018-11-08")
}
]
}
]
},
{
$and: [
{
$eq: [
{
$year: "$creationDate"
},
{
$year: new Date("2018-10-08")
}
]
},
{
$lt: [
{
$month: "$creationDate"
},
{
$month: new Date("2018-11-08")
}
]
},
]
},
{
$and: [
{
$lt: [
{
$year: "$creationDate"
},
{
$year: new Date("2018-11-08")
}
]
},
]
}
],
}
]
}
}
}
])