MongoDB and node.js (Part 3): Looking at Query Results: It’s Complicated

Now that the data is stored in MongoDB (, let’s get it out again through queries and examine the queries’ results.

Before actually running the queries, a little detour is in order. This is about printing an object on the console directly and printing it after applying JSON.stringify(). Here is an example (running in a node.js shell):

> console.log("MongoDB")
> console.log(JSON.stringify("MongoDB"))

Both cases work out fine as they should. So no surprise there.

Let’s try something more interesting:

> console.log(NaN)
> console.log(JSON.stringify(NaN))

Now that’s a bummer. The reason this is a bummer is two-fold:

  • In Part 2 of this blog series we painstakingly made sure that all possible JavaScript types, including the constants, are stored into MongoDB. Retrieving those means that JSON.stringify() must not be in the code path accessing MongoDB in order to not introduce errors as just shown.
  • Furthermore, the underlying BSON implementation of the node.js driver of MongoDB implements a few toJSON() functions that are actually used by JSON.stringify() as it honors overwritten functions.

The first issue can be resolved by writing a separate function that encapsulates JSON.stringify() and implements the “replacer” function (

The second issue means that the node.js driver in MongoDB decided to represent some of the BSON types differently when converting to a string. If not converted, the type is represented differently. So there is an inherent difference between the structure as retrieved from MongoDB and the structure as produced by JSON.stringify().

As a developer you have to decide to avoid JSON.stringify() or to embrace it (meaning, writing a wrapper function that works for all cases).

So, with all that said, the next installment of this blog series will retrieve all data that has been stored in Part 2 and print it out twice, with console.log() as well as console.log(JSON.stringify()).