Mongoose Birthday query - node.js

Question

I store the date of birth (dob) within a user model and I would like to check if it is the users birthday today. How do I do query that?
I have started to try, but obviously failed
// User Model
var UserSchema = new Schema({
name: String,
email: { type: String, lowercase: true },
role: {
type: String,
default: 'user'
},
hashedPassword: String,
dob: { type: Date, default: new Date() },
joinedOn: { type: Date, default: new Date() },
leftOn: { type: Date },
position: { type: String },
provider: String,
salt: String,
google: {},
github: {}
});
// Birtdhays
exports.birthdays = function(req, res, next) {
var todayStart = moment().startOf('day');
var todayEnd = moment().endOf('day');
User
.find()
.where('dob').gt(todayStart).lt(todayEnd)
.limit(3)
.sort('dob')
.select('name dob')
.exec(function(err, users){
if(err) return res.send(500, err);
res.json(200, users);
});
};

Answer 1

Presuming that you have birthdays stored as a date object in your documents then they probably look something like this:
{
"name": "Neil",
"dob": ISODate("1971-09-22T00:00:00Z")
}
So it's not just the "day" but the full year as well from the originally selected day of birth. This probably seems like a logical way to store a date of birth and it is a useful date object for many purposes. But how to query on that? Any date derived from the current year is not going to match that value within a range and other users data on the same day can occur in different years.
There is some JavaScript date manipulation you can do in order to deal with this though, and also some functionality of MongoDB in the aggregation framework ( or alternately using JavaScript in mapReduce ), to get data out of the date that is useful for matching.
Firstly you can look at the $dayOfyear operator and code to get that from the current date to use as a match:
var now = new Date();
var start = new Date( now.getFullYear() + "-01-01" );
var diff = now - start;
var oneDay = 1000 * 60 * 60 * 24;
var dayOfYear = Math.floor(diff / oneDay);
User.aggregate(
[
{ "$project": {
"name": 1,
"dob": 1,
"dayOfYear": { "$dayOfYear": "$dob" }
}},
{ "$match": { "dayOfYear": dayOfYear }
],
function(err,users) {
// work here
}
)
Now that's all fine, or it seems. But of course what happens when there is a leap year? All days after February 28 are moved forward one. You could account for this in other ways, but how about just using the "day" and "month" to do the match on instead:
var now = new Date();
var day = now.getDate(); // funny method name but that's what it is
var month = now.getMonth() + 1; // numbered 0-11
User.aggregate(
[
{ "$project": {
"name": 1,
"dob": 1,
"day": "$dayOfMonth",
"month": "$month"
}},
{ "$match": {
"day": day,
"month": month
}}
],
function(err,users) {
// work here
}
)
And the aggregation operators for $dayOfMonth and $month help there.
So that's all better and you can query for the "birthday" now, but there is something still not really right. Whilst the query will work, it's clearly not really efficient. Since what you are doing here is running through all of the results in the collection and manipulating the existing dates in order to extract the parts to perform a match.
Ideally you don't want to do this, and have the "query" itself target the current "birthdays" in a simple stroke to avoid doing this manipulation to get a match. This is where modelling comes in, and where you should consider adding more data to your document where queries like this are common:
{
"name": "Neil",
"dob": ISODate("1971-09-22T00:00:00Z"),
"day": 22,
"month": 9
}
Then it's easy to query on this as both "day" and "month" fields can also be indexed to further improve performance and avoid scanning the whole collection for matches:
var now = new Date();
var day = now.getDate(); // funny method name but that's what it is
var month = now.getMonth() + 1; // numbered 0-11
User.find({ "day": day, "month": month },function(err,users) {
// work here
})
So those are the considerations. Either accept the manipulation with the aggregation framework ( or possibly mapReduce ), or where you are going to frequently use such a query and/or have many items in the collection then add additional fields to your document schema that can be used as a data point directly in the query itself.

Answer 2

I solved it in the following way:
Added the fields birthday and brithmonth to the user model and added a pre save hook to keep them updated on any changes to the birthday:
UserSchema
.pre('save', function(next) {
this.birthmonth = this.dob.getMonth() + 1;
this.birthday = this.dob.getDate()
next();
});
Thanks Neil for the inspiration!