Using MongoDB

Connect to / Import into a Mock Database

First, let’s connect to the (mock) server and get a handle for our client.

from mongomock import MongoClient
#from pymongo import MongoClient

client = MongoClient()

A MongoDB instance can host multiple databases, which are created dynamically. Here, we will supply a database name as an attribute of the client object, which will prompt MongoDB to create the database with that name if it doesn’t exists.

# Refer to the Software Carpentry ("swc") database
db = client.swc
print(db)

Database(mongomock.MongoClient('localhost', 27017), 'swc')

We see here that what we’re calling db is a Database with the name “swc” that we’re accessing through a Mongo client connected to localhost on port 27017.

print(db.materials)

Collection(Database(mongomock.MongoClient('localhost', 27017), 'swc'), 'materials')

A MongoDB database is organized as a set of collections, each of which contains a set of documents. To first order, you can for now think of a collection as corresponding to a table in SQL and a collection document as corresponding to a table row in SQL.

Just as with databases themselves, database collections are created dynamically in MongoDB. Above, we created a materials collection in our database simply by referring to it by name.

Now, let’s load data from a file and import it as documents into our collection:

import json

We first import the json module from the Python standard library. JSON, which stands for “Javascript Object Notation”, is a way to express simple data structures that is widely used in web-based applications. We’ll go over the format in the next topic when we construct a document to insert into our collection, but for now let’s focus on importing data that we’re given.

with open('data/mongo-novice-materials.json') as f:
    db.materials.insert_many(json.load(f))

We are using a Python context manager to open a file and ensure that it is closed when we are done processing the file contents. In this case, we use the json module to load the file contents as Python-native data structures, which we then hand off to the insert_many method of database collections to insert all of the loaded documents.

db.materials.count()

To confirm we have the data loaded, we use the count() method of a collection object and see that we have more than zero documents in our materials collection.

db.publications.count()