Applying condition to multiple documents for the same field in MongoDB

Question

I have a documents with the following structure:

user {id: 123,  tag:"tag1"}
user {id: 123,  tag:"tag2"}
user {id: 123,  tag:"tag3"}
user {id: 456,  tag:"tag1"}

Given user id I want to find if this user has records with all 3 tags (AND operand). If user has records for "tag1" AND "Tag2" AND "tag3" then return true

SQL equivalent would be like this:

SELECT * FROM users WHERE 
  EXISTS (SELECT * FROM tags WHERE user_id = users.id AND name ='tag1') AND
  EXISTS (SELECT * FROM tags WHERE user_id = users.id AND name ='tag2') AND 
  user_id=123

How can I express something simular in MongoDB?

Answer 1

Because MongoDB does not have a concept of joins, you need to work across this by reducing your input to a single document.

If you change your document structure so that you are storing an array of tags, like the following:

{id: 123,  tag:["tag1","tag2","tag3"]}
{id: 456,  tag:["tag1"]}

You can do a query like the following:

db.user.find({$and:[{tag:"tag1"},{tag:"tag2"},{tag:"tag3"}]})

If needed, you could write a map-reduce to emit an array of tags for a userid, so you can do the above query on it.

Edit - Including a simple map-reduce

Here is a really simple map-reduce to get the initial input into a format that is useful for the find query given above.

var map = function() {emit(this.id, {tag:[this.tag]});}
var reduce = function(key, values){
   var result_array=[];
   values.forEach(function(v1){             
       v1.tag.forEach(function(v2){
        result_array.push(v2);
        });
    });
return {"tag":result_array};}

var op = db.user.mapReduce(map, reduce, {out:"mr_results"})

Then you can query on the map-reduce output collection, like the following:

db.mr_results.find({$and:[{"value.tag":"tag1"},{"value.tag":"tag2"}, {"value.tag":"tag3"}]})