MongoDB Queries in a Nutshell
This wikipage will attempt to get you the 'low hanging fruit' in with MongoDB. With the following commands, you can do 80% of what a typical BH engineer will have to do with their data. If you want further information, please see MongoDB's online documentation: MongoDB Reference Manual.
As you will see, MongoDB's computer language is fairly easy to learn. Whilst it isn't extremely human readable, it is certainly more straightforward than things like nested loops. All of MongoDB's language is based on JSON syntax. It is worth looking a bit at the sytanx before jumping into Mongo Queries. Simply speaking, JSON script will usually look like this:
{"Key": Value} or {"Mongo Query Name": Stuff to Do}
Ok, now let's do some examples. You'll need to do the following:
- Open an instance of MongoShell (Details Here)
- Open Robo3T (Details Here)
The $aggregate function that executes a list of queries one after the other. So let's break that down with what we already know:
Queries = JSON objects Queries = Programmatic Sieves Programmatic Sieves = Filters for us to get to the #DataWeWant
Therefore, by the transitive propery, the aggregate function takes in a list of JSONs and spits out the the subset of data from the database that we care about a.k.a the #DataWeWant.
Aggregate(database,List of JSONs) = #DataWeWant from database
When you use the grasshopper mongo tools, this is what is going on behind the scenes. Recalling that all queries are JSON objects, we should expect to see that the aggregate command takes in a set of JSON objects. To create this list, we will separate each query by a comma:
Aggregate(database,JSON1,JSON2,JSON3,...,JSON9999999) = #DataWeWant from database
Ok, so how do we actually execute this on our Mongo Database? Well, you could use the MongoShell. Or you could use the GrasshopperToolkit tools. But one of the easiest mediums to debug you code is Robo3T. This is because Robot3T has a skosh of intellesense that can help you understand where you have syntax errors in your aggregate command. Recall that if you execute an aggregate command in Robo3T and it works, you can feed the exact same list of queries into Grasshopper using the Mongo Components and it will yield the same result. Therefore, Robo3T is your best friend when you are just getting started with Mongo.
Let's set up our aggregate function. First, we open a collection:
As you can see, when I open my collection, the default command Robot3T throws in is the find() command. If we want to play with $aggregate commands, we need to remove this:
db.getCollection('Model A Forces Database').find({}) --------> db.getCollection('Model A Forces Database')
And in its place add the aggregate command:
db.getCollection('Model A Forces Database').aggregate([JSON1,JSON2])
As you can see, the syntax has a set of [] that enlose the JSON queries. In Robot3T, it will be most convenient to to set up the aggregate command with adding some returns to make the list more human readable:
In what follows, you will learn some common queries that will replace the JSON#'s in the picture above.
THING THAT WILL ANNOY YOU ALERT: To execute any command in Robot3T you must use the F5 key not the ENTER key. This goes against everything you've ever known.....so be prepared to a bit of cussing when instead of Robot3T executing your code, it just returns to the next line of the commandline instead.
If you wish to AUGMENT your data, i.e. add parameters to it, this is your query.
If you wish to GRAB A SUBSET OF YOUR DATABASE based on certain criteria, i.e. filter all data items based a conditional criteria, this is your query.
If you wish to GET ONLY CERTAIN ASPECTS OF YOUR DATA but for all of your database, i.e. I have a million items that all have properties X, Y, and Z, but I only care out the X and Y values, this is your query.
{$project: {FX: 1}}
If you wish to TRANSFORM THE STRUCTURE OF YOUR DATA, i.e. instead of having 10,000 lists of 2 items, having 2 lists of 10,000 items, this is your query.