Storing and querying time intervals in MongoDB - mongodb

Question

Modeling shift planning app:
I came up with such data structure describing the shift.
{
"fromHour" : 7,
"fromMinute" : 30,
"toHour" : 9,
"toMinute" : 30,
"week" : 5,
"date" : "2015-01-26",
"user" : {
// ...
},
"_id" : ObjectId("54d0e4a82b9dc26c0c0f36e7")
}
The main thing that I need to store is the information:
1. when shift starts (hour and minute),
2. when shift ends (hour and minute) and
3. what date it's actually happening.
The fields fromHour, fromMinute, toHour, toMinute and date as ISO string worked pretty good for me for storing and querying the shifts by particular date.
The problem occurred when I needed to build reports out of it. Say, I want to get all shifts from "2015-01-01" to "2015-02-01" in range from 07:00, till 23:00.
I can add $and clause to my query, like
[ { fromHour: { '$gte': 7 } },
{ fromMinute: { '$gte': 0 } },
{ toHour: { '$lt': 23 } },
{ toMinute: { '$lt': 0 } } ]
But that doesn't work good, since for shift there toMinute is 30 the $lt will be false.
I'm trying to find efficient data structure that would allow to store timespans that would be easy to query.

Answer 1

Storing hours and minutes in two different fields separated is too error-prone and makes your job harder. Since Mongo does not have a distinct "Time" data type, only Date, and shifts usually start and end at "easy" times, I would recommend to implement something like converting the time to a real number in your application like this:
00:00 --> 0
01:00 --> 1
...
08:00 --> 8
08:15 --> 8.25
08:30 --> 8.5
...
16:30 --> 16.5
...
It is a bit of extra work in the app because you have to convert while saving or displaying but it's still better than having one single time value in two different fields.
So your data would look like this:
{
"shiftStart" : 7.5,
"shiftEnd" : 9.5,
"week" : 5,
"date" : "2015-01-26",
"user" : {
// ...
},
"_id" : ObjectId("54d0e4a82b9dc26c0c0f36e7")
}
and your query:
[ { shiftStart: { '$gte': 7 } },
{ shiftEnd: { '$lt': 23 } } ]

Answer 2

You can store the data with the date type in ISODate(...) format and then use $project and Date Aggregation Operators to query the data.
For your example:
db.shifts.aggregate([
{ $match: //matches the dates first to filter out before $project step
{ datetimeStart:
{ $gte: ISODate("2015-01-01T07:00:00.000Z"),
$lt: ISODate("2015-02-01T00:00:00.000Z")
},
datetimeEnd:
{ $gte: ISODate("2015-01-01T07:00:00.000Z"),
$lt: ISODate("2015-02-01T00:00:00.000Z")
}
}
},
{ $project: // $project step extracts the hours
{ otherNeededFields: 1, // any other fields you want to see
datetimeStart: 1,
datetimeEnd: 1,
hourStart: { $hour: "$datetimeStart" },
hourEnd: { $hour: "$datetimeEnd" }
}
},
{ $match: // match the shift hours
{ hourStart: { $gte: 7 },
hourEnd: { $lte: 23 }
}
}
])
With this system it would be possible, but complicated to find something more like shifts between 7:30AM and 10:30PM:
db.shifts.aggregate([
{ $match: //matches the dates first to filter out before $project step
{ datetimeStart:
{ $gte: ISODate("2015-01-01T07:30:00.000Z"),
$lt: ISODate("2015-02-01T00:00:00.000Z")
},
datetimeEnd:
{ $gte: ISODate("2015-01-01T07:30:00.000Z"),
$lt: ISODate("2015-02-01T00:00:00.000Z")
}
}
},
{ $project: // $project step extracts the hours
{ otherNeededFields: 1, // any other fields you want to see
datetimeStart: 1,
datetimeEnd: 1,
hourStart: { $hour: "$datetimeStart" },
minStart: { $minute: "$datetimeStart" },
hourEnd: { $hour: "$datetimeEnd" },
minEnd: { $minute: "$date
}
},
{ $match: // match the shift hours
{ $or:
[
{hourStart: 7, minStart: {$gte: 30}}, // hour is 7, minute >= 30
{hourStart: { $gte: 8 }} // hour is >= 8
],
$or:
[
{hourEnd: 22, minEnd: {$lte: 30}}, // hour is 22, minute <= 30
{hourEnd: { $lte: 21 }} // hour is <= 21
]
}
}
])