How can I write a Mongoose find query that uses another field as it's conditional? - mongodb

Question

Consider the following:
I have a Mongoose model called 'Person'. In the schema for the Person mode, each Person has two fields: 'children' and 'maximum_children'. Both fields are of type Number.
I would like to write a find query that returns Persons when that Persons 'children' value is less that it's 'maximum_children' value.
I have tried:
person_model.find({
children: {
$lt: maximum_children
}
}, function (error, persons) {
// DO SOMETHING ELSE
});
and
person_model.find({
children: {
$lt: 'maximum_children'
}
}, function (error, persons) {
// DO SOMETHING ELSE
});
I'm doing something wrong in trying to specify the field name that I want to compare 'children' against.

Answer 1

OK.
I found a solution, just after I posted this question.
The answer seems to be:
person_model.find({
$where: "children < maximum_children"}, function (error, persons)
}, {
// DO SOMETHING ELSE
});
Seems to work OK, although it seems messy.

Answer 2

$where must execute its JavaScript conditional against every doc so its performance can be quite poor. Instead, you can use aggregate to include a new field in a $project stage the indicates whether the doc matches or not and then filter on that:
person_model.aggregate([
{$project: {
isMatch: {$lt: ['$children', '$maximum_children']},
doc: '$$ROOT'
}},
{$match: {isMatch: true}},
{$project: {_id: 0, doc: 1}}
], function(err, results) {...});
This uses $$ROOT to include the original doc as the doc field of the projection, with a final $project used to remove the isMatch field that was added.
results looks like:
{
"doc" : {
"_id" : ObjectId("54d04591257efd80c6965ada"),
"children" : 5,
"maximum_children" : 10
}
},
{
"doc" : {
"_id" : ObjectId("54d04591257efd80c6965add"),
"children" : 5,
"maximum_children" : 6
}
}
If you want to remove the added doc level of the objects you can use Array#map on results like so:
results = results.map(function(item) { return item.doc; });
Which reshapes results to put them back into their original form:
{
"_id" : ObjectId("54d04591257efd80c6965ada"),
"children" : 5,
"maximum_children" : 10
},
{
"_id" : ObjectId("54d04591257efd80c6965add"),
"children" : 5,
"maximum_children" : 6
}