An Introduction to NoSQL Databases


No SQL Illustration1

This article, which is an introduction to NoSQL databases, takes the reader through structured query languages and then on to MongoDB. It touches upon the integration of Python with MongoDB.

With the growth and development of cloud computing technology there has been a corresponding rise in NoSQL databases. In order to understand NoSQL databases, we have to study the need for them. Before one heard about NoSQL, there were many types of databases and related technologies. Most of these were relational databases, and to work on these effectively, SQL or structured query languages were developed. But these databases, as well as SQL, posed complex problems, for which developers tried to find alternate solutions. Some major problems were rigid schema designs, joins, child-parent dependencies, and so on. Developers had to look for alternatives that would ease these common problems with relational databases and SQL. Hence, ‘Not only SQL’ or NoSQL came into existence, which offered support for SQL-like query languages, besides other features. Some of its key features are:

  • Dynamic schema design or non-schema
  • •No dependency (non-relational model)
  • Flexibility

The following are the types of NoSQL databases (DBs) commonly found in today’s market:

  • Key-value databases
  • •Document databases
  • •Column family store databases
  • •Graph databases

In this tutorial, we will start looking into different databases, beginning with a document based DB called MongoDB, which is quite popular among developers worldwide.

This is a document based NoSQL DB. It has become quite popular in recent years and many big organisations are major supporters of it. It was initially launched in 2009, and is developed and distributed by the 10gen organisation. It is open source and is freely available for developers to explore and learn.

MongoDB set-up and installation
MongoDB comes in both 32-bit and 64-bit versions for different OS platforms. You can get the installation details from the MongoDB site.

Note: It is always preferable to install the 64-bit version for better performance.

Installation on Ubuntu
Installation on Ubuntu is very easy unlike the complexity with a Windows machine. If you want a simple installation, then follow the steps mentioned below.
Run the following commands from a terminal:

$ sudo apt-get update
$ sudo apt-get install Mongoloid install –y mongodb-org

Start the mongod service by executing the commands shown below:

$ sudo service mongod start

Open another terminal window and execute the following command:

$ mongo

The above command opens the shell prompt, which will connect to MongoDB.
Your MongoDB installation is now complete and ready to be used.
Sometimes, the above method may not work for Debian Linux. Don’t be disheartened; download the .gz file of MongoDB. For me, it is mongodb-linux-x86_64-debian71-3.0.1.gz. Extract it by using the following command:

# tar –xvzf mongodb-linux-x86_64-debian71-3.0.1.gz

After that, run the command shown below:

export PATH=$PATH:/mongodb/mongodb-linux-x86_64-debian71-3.0.1/bin

The above command adds the MongoDB binary directory to the Linux path.
Run the following commands to run the service:

# cd /mongodb/mongodb-linux-x86_64-debian71-3.0.1/bin
# ./mongod --dbpath=/data/mongodb3

A better way to do the above task is to put both commands in one file saved as, then run the shell script.
For the Mongo terminal, run the following commands:

# cd /mongodb/mongodb-linux-x86_64-debian71-3.0.1/bin
# ./mongo

You can save both commands in a file saved as, and then run the shell script.
You can install MongoDB in Windows too. Check the steps of the installation under the NoSQL tab of

MongoDB with Python
Now let’s study how Python and MongoDB interact. You can run MongoDB and Python programs on Linux. In order to install the pymongo module, you can take advantage of the PIP program. Use the following command:

# pip install pymongo

After entering the command, the progress of the installation will be displayed.
Now let’s explore some Python code for MongoDB.
Insert data with Python. MongoDB stores the data in the JSON format. I am presenting a JSON file named test.json for testing:

“Records”: [
“record”: “from socrates to sartre: the philosophic quest (paperback)”,
“ID”: 39
“record”: “the magic of believing by bristol claude m.-english-embassy books-paperback (paperback)”,
“ID”: 40
“record”: “haunted: hardy boys (paperback)”,
“ID”: 42
“record”: “the intelligent investor (paperback)”,
“ID”: 44
“record”: “understanding & worshiping sri chakra (paperback)”,
“ID”: 45
“record”: “pmr: i too had a love story (paperback)”,
“ID”: 55


In the above JSON file, the record field represents the actual record and the ID field represents the number (record number).
Now, let’s insert the test.json file in MongoDB using Python:

import json
import pymongo
connection = pymongo.MongoClient(“mongodb://localhost”)
record1 = db.bookcollection
page = open(“test.json”, ‘r’)
parsed = json.loads(

for item in parsed[“Records”]:

Run the code using the following command:

# python

Just be sure that test.json is present in the same directory.
Now, let’s go over each and every line.
The first two lines are importing the json and pymongo module.
Code line connection = pymongo.MongoClient (“mongodb://localhost”) establishes a connection to the database. The next line is handle to the book database. Code line record1 = db.bookcollection; here, record1 contains a collection of records. In the next line, page = open (“test.json”, ‘r’) the variable page contains the test.json file. And in the following line, the variable parsed contains the file in json format. If you want to view the records from the MongoDB command line, use the commands as shown below.
The code line for the item in parsed[“Records”]: shows that we are interested in parsed[“Records”]. The last line inserts the records one by one in record1 collection.
If you want to view the records from the MongoDB command line, use the following command:

# use book
# db.bookcollection.find()

Find the data with Python
In the previous code snippet, we inserted the records in the record1 collection of books database. Now we are going to discuss the code to view the records.

Here is the minimum code named to view the records:

import pymongo
connection = pymongo.MongoClient(“mongodb://localhost”)
record1 = db.bookcollection
cursor = record1.find()
for doc in cursor:
print doc
Figure 1: Output shows the records of bookcollection

I think you are acquainted with the first four lines. Code line cursor = record1.find() tells the cursor to contain all the records. If you print the cursor variable, it will give you addresses like <pymongo.cursor.Cursor object at 0x1aac810>
In order to print the cursor document, we have used the for loop.
Run the program and you will get the output shown in Figure 1.
Let’s assume that you want to display a particular field. Make some amendments to the above code. Replace the following line…

cursor = record1.find()


cursor = mo.find({},{«ID»:True,

…and run the program again. The output will be like what’s shown in Figure 2.
Here, ID is a particular field.

Figure 2: Output shows the specific field

Now let’s assume that you want to display a particular type, say, ID= 39. Make amendments to the above program and replace the following line…

cursor = mo.find({},{“ID”:True, “_id”:False}) with cursor = record1.find({“ID”:39},{“ID”:True, “_id”:False})

…and run the program again. The output will be like what is shown in Figure 3.

Figure 3: Output shows the specific document (record)

Remove records using Python
You can remove a particular record using the pymongo module. I am going to give the minimum required code to delete the record.

import pymongo
connection = pymongo.MongoClient(“mongodb://localhost”)
record1 = db.bookcollection

result = record1.remove({‘ID’:55})
print “num removed: “, result

Run the above program and the output will be as shown in Figure 4.

Figure 4: Output shows document(record) is removed

Update records using Python
Python can be used to update records as well as add new fields in a particular record.
In the first update program, let’s update an old value. Let us discuss our next program named

import pymongoconnection = pymongo.MongoClient(“mongodb://localhost”)db=connection.bookrecord1 = db.bookcollection
result = record1.update({‘ID’:45},{‘$set’:{‘ID’:100}})
print result
Figure 5: Output of the program that updates an old field
Figure 6: Output of the program that adds a new field

Run the above program and the output will be as shown in Figure 5.

Code line record1.update({‘ID’:45},{‘$set’:{‘ID’:100}}) is responsible for updating the record ID=45 to ID= 100. You can check with the help of find programs.
Consider a situation in which you have to add one more field. Now, I am going to present the next code named

import pymongo
import datetime
connection = pymongo.MongoClient(“mongodb://localhost”)
record1 = db.bookcollection

result = record1.update({‘ID’:100},{‘$set’:{‘website’:””}})

print record1.find_one({‘ID’:100})

The above code adds one more field with the value website’:” in the record, whose ID =100.
Run the program The output will be as shown in Figure 6.


Please enter your comment!
Please enter your name here