titanic_mongo.py

from pymongo import MongoClient

#########################
Establishing a connection 
#########################
client = MongoClient("localhost", 27017) #connects to localhost by default
db = client["titanic"]

###############################
Verifying connection to MongoDB
###############################
print("Total number of documents in the 'titanic' collection: ", db.guests.count_documents({}))
print(db.guests.find_one())

##########
#Filtering
##########
print("\n----------------Filtering--------------------")

The most compelling question when it comes to Titanic - How many survivors?
print(db.guests.count_documents({"survived":1}))
print(db.guests.count_documents({"age":{"$lt":18}}))


print(db.guests.count_documents({"$and":[{"survived":1}, {"age":{"$lt":18}}]}))

print(db.guests.count_documents({"name":{"$regex":"Mis"}}))

#Unlike relational databases, presence isn't compulsory in MongoDB. 
#There's a filter to check for presence.
#Do all passengers have a ticket_number?
print(db.titanic.count_documents({"ticket_number":{"$exists":False}}))

################
#Distinct values
################
print("\n--------------Distinct Values----------------")

print(db.titanic.distinct("point_of_embarkation"))
print(db.titanic.distinct("class", {"fare_paid":0}))

############
#Projections
############
print("\n----------------Projections------------------")

Counting documents is straightforward but on using the find() method, a cursor is returned.
print(db.titanic.find({"parents_children":{"$gte":5}}))

We can send the cursor over to a list and slice the list to limit the results
print(list(db.guests.find({"parents_children":{"$gte":5}}))[:3])

This cursor can also be iterated over using a loop
from pprint import pprint
for doc in db.guests.find({"parents_children":{"$gte":5}}):
    pprint(doc)

#How do we limit the fields that we see in the results?
for doc in db.titanic.find({"parents_children":{"$gte":5}}, {"name":1, "age":1, "_id":0}):
    print(doc)

If we aren't excluding any fields in our projections, pymongo accepts another simpler format.
for doc in db.titanic.find({"parents_children":{"$gte":5}}, ["name", "age"]):
    print(doc)

#Writing the above code in a single line using list comprehension
[print(doc) for doc in db.titanic.find({"parents_children":{"$gte":5}}, ["name", "age"])]

#Python's use in dealing with iterables like cursors and MongoDB's projection can be used to
#slim down the documents returned. Togther, they make for powerful querying!

########
#Sorting
########
print("\n----------------Sorting------------------")
for doc in db.guests.find({"age":{"$lt":18}}, sort = [("survived", 1)]):
    print(doc["class"],doc["survived"])

for doc in db.titanic.find({"age":{"$lt":18}}, sort = [("class", 1), ("gender", 1)]):
    print(doc["class"], doc["gender"], doc["survived"])
    
#Sorting in the Mongo shell is slightly different. Instead of tuples, they have to be entered 
#as key-value pairs. Eg: {"class":1, "gender":1}    

######################
#Limiting and Skipping  
######################
print("\n------------Limiting & Skipping---------------")

for doc in db.titanic.find({"parents_children":{"$gte":5}}, skip = 3, limit = 3):
    print("{name}: {survived}".format(**doc))

#####################
#Aggregation Pipeline
#####################
print("\n------------Aggregation Pipeline---------------")

print(db.titanic.count_documents({"age":{"$gte":70}}))

from collections import OrderedDict
cursor = db.titanic.aggregate([ \
    {"$match":{"age":{"$gte":70}}}, \
    {"$project":{"class":1, "gender":1, "survived":1}}, \
    {"$sort":OrderedDict([("class", 1), ("gender", 1)])}, \
    {"$limit":5}])
for doc in cursor:
    print(doc["class"], doc["gender"], doc["survived"])