Complex MongoDB query aggregate - mongodb

Question

I've a document representing a house embedding many rooms. The structure is this.
{
description: 'House'
total_price: 1000,
rooms: [{
description: 'Small Single Room',
type: single,
available: true,
price: 300
}, {
description: 'Big Single Room',
type: single,
price: 400,
available: true
}, {
description: 'Another Room',
type: single,
price: 300,
available: true
}]
}
I'm having hard time to solve this question: how can I get all houses having available single rooms for a total price of 700.
To do this I need what follows.
set a condition on some embedded document properties
set a condition on the sum of the available embedded rooms
set a condition on the sum
At the moment I'm trying to understand how this can be done. I've been looking to aggregate and other methods, but I couldn't find a working solution.
Thanks a lot.

Answer 1

Here is one approach that you could use. Save remaining rooms and remaining price at root level (initialize with total price and total rooms). Whenever you set available as false for a room, subtract the price and count from remaining_price and remaining_rooms.
{
description: 'House'
total_price: 1000,
remaining_price: 700,
remaining_rooms: 2,
rooms: [{
description: 'Small Single Room',
type: single,
available: false,
price: 300
}, {
description: 'Big Single Room',
type: single,
price: 400,
available: true
}, {
description: 'Another Room',
type: single,
price: 300,
available: true
}]
}
Now you can simply query the rooms you want using (I am writing the query using mongoid, which you can convert to whatever you are using):
House.where(:remaining_price.lte => 700, :remaining_rooms.gte => 2).to_a

Answer 2

Here's an aggregation framework answer.
Note that the first $match filters documents at the collection level
and uses $elemMatch for matching both type and availability on an array element.
In order to do the sum of the prices of the available rooms for a house,
the unavailable rooms are necessarily filtered out and lost.
Addition fields in the $group stage preserve data for viewing,
but it may be more appropriate for your app to fetch full docs via the _id's.
Hope that this answers your question in its original intent
and demonstrates the power of the aggregation framework.
test/unit/house_test.rb
require 'test_helper'
require 'pp'
class HouseTest < ActiveSupport::TestCase
def setup
House.delete_all
end
test "complex aggregate query" do
puts "\nMongoid::VERSION:#{Mongoid::VERSION}\nMoped::VERSION:#{Moped::VERSION}"
total_price_limit = 700
pipeline = [
{'$match' => {'rooms' => {'$elemMatch' => {'type' => 'single', 'available' => true}}}},
{'$unwind' => "$rooms"},
{'$match' => {'rooms.available' => true}},
{'$group' => {
'_id' => '$_id',
'description' => {'$last' => '$description'},
'total_price_available' => {'$sum' => '$rooms.price'},
'rooms' => {'$push' => '$rooms'},
}
},
{'$match' => {'total_price_available' => {'$lte' => total_price_limit}}},
{'$sort' => {'total_price_available' => 1}},
]
docs = [
{
description: 'House 1',
total_price: 1000,
rooms: [{
description: 'Small Single Room',
type: 'single',
available: true,
price: 300
}, {
description: 'Big Single Room',
type: 'single',
price: 400,
available: true
}, {
description: 'Another Room',
type: 'single',
price: 300,
available: true
}]
},
{
description: 'House 2',
total_price: 600,
rooms: [{
description: 'Small Single Room',
type: 'single',
available: true,
price: 300
}, {
description: 'Big Single Room',
type: 'single',
price: 400,
available: false
}, {
description: 'Another Room',
type: 'single',
price: 300,
available: true
}]
}
]
docs.each { |house| House.create(house) }
pp House.collection.aggregate(pipeline)
end
end
$ rake test
Run options:
# Running tests:
[1/1] HouseTest#test_complex_aggregate_query
Mongoid::VERSION:3.1.5
Moped::VERSION:1.5.1
[{"_id"=>"5272dbe2e4d30b7e0a000002",
"description"=>"House 2",
"total_price_available"=>600,
"rooms"=>
[{"description"=>"Small Single Room",
"type"=>"single",
"available"=>true,
"price"=>300},
{"description"=>"Another Room",
"type"=>"single",
"price"=>300,
"available"=>true}]}]
Finished tests in 0.046359s, 21.5708 tests/s, 0.0000 assertions/s.
1 tests, 0 assertions, 0 failures, 0 errors, 0 skips