Due to the nature of MongoDB, it's perfectly acceptable to have different data type objects in the same field. This may happen by accident or on purpose (that is, null and actual values in a field).

The sorting order of different types of data, from highest to lowest, is as follows: 

1. Max key (internal type)
2. Regular expression
3. Timestamp
4. Date
5. Boolean
6. ObjectID
7. Binary data
8. Array
9. Object
10. Symbol, string
11. Numbers (int, long, double)
12. Null
13. Min key (internal type)

Non-existent fields get sorted as if they have null in the respective field. Comparing arrays is a bit more complex than fields. An ascending order of comparison (or <) will compare the smallest element of each array. A descending order of comparison (or >) will compare the largest element of each array.

For example, see the following scenario:

> db.types.find()
{ "_id" : ObjectId("5908d58455454e2de6519c49"), "a" : [ 1, 2, 3 ] }
{ "_id" : ObjectId("5908d59d55454e2de6519c4a"), "a" : [ 2, 5 ] }

In ascending order, this is as follows:

> db.types.find().sort({a:1})
{ "_id" : ObjectId("5908d58455454e2de6519c49"), "a" : [ 1, 2, 3 ] }
{ "_id" : ObjectId("5908d59d55454e2de6519c4a"), "a" : [ 2, 5 ] }

However, in descending order, it is as follows:

> db.types.find().sort({a:-1})
{ "_id" : ObjectId("5908d59d55454e2de6519c4a"), "a" : [ 2, 5 ] }
{ "_id" : ObjectId("5908d58455454e2de6519c49"), "a" : [ 1, 2, 3 ] }

The same applies when comparing an array with a single number value, as illustrated in the following example. Inserting a new document with an integer value of 4 is done as follows:

> db.types.insert({"a":4})
WriteResult({ "nInserted" : 1 })

The following example shows the code snippet for a descending sort:

> db.types.find().sort({a:-1})
{ "_id" : ObjectId("5908d59d55454e2de6519c4a"), "a" : [ 2, 5 ] }
{ "_id" : ObjectId("5908d73c55454e2de6519c4c"), "a" : 4 }
{ "_id" : ObjectId("5908d58455454e2de6519c49"), "a" : [ 1, 2, 3 ] }

And the following example is the code snippet for an ascending sort:

> db.types.find().sort({a:1})
{ "_id" : ObjectId("5908d58455454e2de6519c49"), "a" : [ 1, 2, 3 ] }
{ "_id" : ObjectId("5908d59d55454e2de6519c4a"), "a" : [ 2, 5 ] }
{ "_id" : ObjectId("5908d73c55454e2de6519c4c"), "a" : 4 }

In each case, we highlighted the values being compared in bold. 

We will learn about the data type in the following section.