Refrence same Document in Multiple Websites - mongodb

Question

I have one Admin interface website where a user can create news articles and also select which websites this article shall appear on. (Have many websites connected to the same Mongo database)
Each website has an array with article ID's. When going to one of these websites I loop this array and fetches all articles (from the Articles collection) belonging to this site with:
Articles.findOne({_id:id});
However this becomes a problem if I like to do more advanced queries such as sorting on date. Putting a limit etc etc.
At the same time I don't want to filter all articles for a specific site directly from the Articles collection since it seams expensive? (it contains all articles from all websites) And saving each article locally on each website would create duplicates.
I wonder what is a good way of storing these news articles and still fetch them quickly for each website?
------------------------
I am currently doing like this for fetching all articles from a site and sorting on date. But now I also need to put a limit, only fetch articles from a specific category and so on and it becomes very bothering:
var websites = Websites.find({name : "SITENAME"},{}).fetch();
var now = new Date();
var articles = [];
websites[0].articles.forEach(function(id) {
article = Articles.findOne({_id:id});
if (article != undefined && article.publishedDate < now) {
articles.push(article);
}
});
articles.sort(function(a, b) {
a = a.publishedDate;
b = b.publishedDate;
return a > b ? -1 : a<b ? 1 : 0;
});
return articles;
Edit to clarify:
This is the current database structure. Each article in the articles collection looks like this:
{
"_id" : "CdHWxgq75yjcgQoDZ",
"category" : "Nyheter",
"tags" : [
"ZaifTyGGouPwdrGur"
],
"data" : [
"Hello this is some random content"
],
"publishedOnSites" : [
"ZaifTyGGouPwdrGur"
],
"publishedDate" : ISODate("2015-11-20T07:22:09.799Z"),
"userId" : "B3t6QFgG7MfNkvzR5"
}
Each website in the websites collection like this:
{
"_id" : "ZaifTyGGouPwdrGur",
"name" : "SITENAME",
"categories" : [
"News",
"Life",
"TV",
"Sport",
"Quizzes",
"Video"
],
"tags" : [
"batman",
"bil",
"polis",
"flicka",
"cool",
"byrå",
"förvandling"
],
"articles" : [
"PgGetxkC9KynaPNLc",
"ZaifTyGGouPwdrGur",
"oPQHh3u2CGhRwYp2a",
"a5ZkhbxRcLEpggTuF",
"t3n8Zp6Cve6e88Gmt",
"eYQmaavt6tAwbbmzf",
"F9LzZFcFxSpejseHn",
"NLWb5NahoPjgAt7eN",
"pwkTtFN8gZCsnKDGg",
"o62uCK7S6qauJfyYa",
"pivJGzo4CFw3QRb3v",
"H2EHv7rX5GQmyqiDk",
"tGfrv82NMwJEpuThK",
"CvjGPKmsCqmd9o5oP",
"29hoZxnmfovTnC8TM",
"NXHXhaXDYgKLagamJ",
"9EjfABeK5akDLeZJT",
"5q5zeYRkPHMJXtEpT",
"eWGwWq3J7JqtQi2fK",
"7W27ufZ4qDyX4mJnC",
"oBhGpNCBTrMcb3qvq",
"7pRorBYbZ8Mx6jYX3",
"d2PoAFGTcbQzapXpW",
"qDRiB65vcpMu6KTTe",
]
}
I save the article ID's in each website to fetch it quickly without having to filter all articles. However this becomes a problem when I want to make queries such as sorting on date, putting a limit, skipping the first elements, only fetch articles with a certain category etc.
I need suggestions for a better database structure.

Answer 1

Usually, it's better to let MongoDB handle the filtering, sorting etc. It knows how to do it well and how to do it fast.
So, what you'd want to do is this:
var arcticles_ids = Websites.findOne({name: "SITENAME"}).articles;
var articlesCursor = Articles.find({_id: {$in: articles_ids}}, {sort: {publishedDate: -1}});
On the second line, you can add a limit etc. If you're worried about performance, add indexes, e.g.:
db.articles.createIndex({_id: 1, publishedDate: -1});
Note: Do not just add this index to your database. Analyze what kind of queries you have and add indexes based off of that. The above was just an example.
Also, you might want to consider adding a field to the Articles collection that stores all the websites that this article belongs to. E.g:
article: {
someField: someValue,
websites_ids: [1, 5, 8, 10]
}
This is useful if you want to make your query reactive. E.g:
var articlesCursor = Articles.find({websites_ids: website_id}, {sort: {publishedDate: -1}});
This way, if the cursor is reactive and an article is added to a website, the client immediately receives this information about the article. If done your way, the cursor would only track the specific IDs of the articles. Something to consider.