LESSON

Non-Relational Database
Types
Google slide deck available here

This work is licensed under the Creative Commons

Attribution-NonCommercial-ShareAlike 3.0 Unported
License
(CC BY-NC-SA 3.0)
 Non-Relational Database Types

Key/Value Graph Column Document

There are four main types of non-relational databases: key/value, graph, column, and
document, and we’ll investigate each in this lesson.
Structure

● A unique key is paired with a

collection of values, where the
values can be anything from a
string to a large binary object.

Strength
Key/Value Database
● Simple data model.

Key-value databases use a very simple schema: a unique key is paired with a
collection of values, where the values can be anything from a string to a large binary
object.

One way that databases using this structure gain in performance is that there are no
complex queries. The system knows on which server the data is located and sends
the request to just that server.

Example: Redis is one of the most popular examples of a key/value store database. .
 Key/Value: Example

Key Value

Name Sherlock Holmes

Age 40

Address 221B Baker Street

As the simplest of the non-relational databases, key/value is exactly as it sounds,

data is organized based on an identifying key and its corresponding value. This
simplicity makes it beneficial for large datasets, but not when complex relationships
are at play.
Structure
● Captures connected data.
● Each element is stored as a
node.
● Connections between nodes
are called links or relationships.

Strength

● Traverses the connections

between data rapidly.
Graph Database

Graph databases are another type within the non-relational family of databases. A
popular example of a graph database being used today is Neo4J.

They have been designed to deal with problems around relationships with and
focuses on connected data. This type of model does capture the richness of the
relationships, however many problems are not naturally modelled as connected data
or relationship problems. This makes this database well matched to these problems
but as these are niche, it does not have a wider or broader applicability to other
problems.

This databases stores information as a collection of nodes and edges, where the
edges represent the relationships between the nodes.

Storing the relationships between data means that related data can often be retrieved
in a single operation. The concept of relationships between data and this
interconnectedness is the key principle behind this type of non-relational database.
This approach counters the approach required in SQL/Relational Databases where many
joins on several attributes in a number of tables would often be required to retrieve these
kinds of relations in the data.

The most popular example of a graph database is Neo4J.

 Graph: Example
221B
BakerS :addre
ss
t.

:li
ke
d
Sherlock
Holmes
ress

s
Post nd
e :fri
1 :fri h wi ends
:add

:p
os wit th
te
d

:friends with
John Irene
Watson Adler
:p
ed

os

:p
st

te

os
d
:like
:po

te
d
Post Post
2 Post
3 1

This is a rudimentary example of how data could be stored in a graph database. As

you will see, this type of store is useful for social media applications where multiple
objects will have multiple relationships or links.
Structure

● Data is stored using key rows

that can be associated with
one or more dynamic
columns

Strengths

● Highly performant queries

● Designed for analytics Column Oriented
or Wide Column

A column oriented or wide column non-relational database is primarily designed for

analytics. Today, Cassandra is a commonly used column oriented database. The
advantage of column versus record/row oriented databases is that column oriented
databases return data in columns making the query much more performant as it will
not return many irrelevant fields that are not required for the query being serviced.
The primary key in a column oriented database is the data / value which is then
mapped to row keys. This is the inverse, or opposite, of how the primary key works in
a relational database.

The structure of the column data is flexible and can vary from row to row. Associated
with storing large amounts of data: billions of rows with millions of columns

This does not necessarily require a separate database and can be implemented as
indexes on existing data structures to add this type of functionality to a database.
 Column Oriented Example

Name ID Age ID Height ID

Sherlock 001 40 001 6’2 001

John 002 45 002 5’9 002

Irene 003 43 003 5’7 003

Here we see an example of what data might look like in a column-oriented database.
Though at first glance it might seem similar to the known relational tabular format, it is
very different in that the data is sorted via column IDs. Therefore the relationships
between data are identified via the column key.
Structure
● Polymorphic data models

● Each document contains markup

that identifies fields and values.

Strengths
● Obvious relationships using
embedded arrays and
documents
Document Database
● No complex mapping

Document databases, such as MongoDB, store data in a single document which can
have different shapes within the single collection or table that stores the documents.

It provides a clear means of capturing relationship using sub-documents and

embedded arrays within a single document.

The document is a close analogy to the object in object oriented programming and
provides a clear natural representation of a ‘thing’ and it’s data.

This clear representation often means that there is no requirement for object mapping
between the database and the application/programming language. The document is
often the exact match for the object that the programmer wishes to use. The flexibility
of the document to hold many shapes or multiple parallel schemas at any point in time
gives great flexibility for modeling with documents when compared to relational
database tables.
 Document Model Example

{
"_id": {
ObjectId("5ef2d4b45b7f11b6d7a" "_id":
), ObjectId("6ef8d4b32c9f12b6d4a")
"user_id": "Sherlock ,
Holmes", "user_id": "John Watson",
"age": 40, "age": 45,
"address": "address":
{ {
“Country: “England” “Country: “England”
“City”: “London”, “City”: “London”,
“Street”: “221B Baker “Street”: “221B Baker
St.” St.”
}, },
“Hobbies”:[ violin, “Medical license”: “Active”
crime-solving ] }
}

The key thing to understand about the document model is that data that is accessed
together is stored together. It is also important to note that just because one
document has one field does not mean another related document has to have the
same field when stored together. We will discuss this more in the lesson when we talk
about collections.
The Document Model

For general purpose use, the

document model prevails as
the preferred model by
developers and database
administrators.

Due to the fact that the document model implements data structures using
programming languages, it is the preferred model by developers as it most closely
matches the way they think and work already.

We will take a closer look at the document model and understand how it functions.
Quiz
 Quiz

Which of the following are types of non-relational databases? More

than one answer choice can be correct.
A. Key/Value

B. Graph

C. Column-oriented or wide-column

D. Document

E. Data Lake
 Quiz
Which of the following are types of non-relational databases? More
than one answer choice can be correct.

A. Key/Value

B. Graph

C. Column-oriented or wide-column

D. Document

E. Data Lake

CORRECT: Key/Value - A key/value database is regarded as a non-relational

database
CORRECT: Graph - A graph database is regarded as a non-relational database
CORRECT: Column-oriented or wide-column - A column-oriented or wide-column
database is regarded as a non-relational database
CORRECT: Document - A document database is regarded as a non-relational
database
INCORRECT: Data Lake - A data lake is regarded as a centralised repository of data,
both structured and unstructured so it not regarded as non-relation or as relational.
 Quiz

Which of the following are types of non-relational databases? More

than one answer choice can be correct.
A. Key-Value A key-value database is
regarded as a
B. Graph non-relational database.

C. Column-oriented or wide-column

D. Document

E. Data Lake

CORRECT: Key-Value - A key-value database is regarded as a non-relational

database
 Quiz

Which of the following are types of non-relational databases? More

than one answer choice can be correct.
A. Key-Value A graph database is
regarded as a
B. Graph non-relational database.

C. Column-oriented or wide-column

D. Document

E. Data Lake

CORRECT: Graph - A graph database is regarded as a non-relational database

 Quiz

Which of the following are types of non-relational databases? More

than one answer choice can be correct.
A. Key-Value A column-oriented or
wide-column database is
B. Graph regarded as a
non-relational database.
C. Column-oriented or wide-column

D. Document

E. Data Lake

CORRECT: Column-oriented or wide-column - A column-oriented or wide-column

database is regarded as a non-relational database
 Quiz

Which of the following are types of non-relational databases? More

than one answer choice can be correct.
A. Key-Value A document database is
regarded as a
B. Graph non-relational database.

C. Column-oriented or wide-column

D. Document

E. Data Lake

CORRECT: Document - A document database is regarded as a non-relational

database
 Quiz

Which of the following are types of non-relational databases? More

than one answer choice can be correct.
A. Key-Value A data lake is regarded
as a centralised
B. Graph repository of data, both
structured and
C. Column-oriented or wide-column unstructured so it is not
regarded as
D. Document
non-relational or as
E. Data Lake relational.

INCORRECT: Data Lake - A data lake is regarded as a centralised repository of data,

both structured and unstructured so it is not regarded as non-relational or as
relational.
Continue Learning! Github Student
Developer Pack

MongoDB University has free self-paced Sign up for the MongoDB Student Pack to
courses and labs ranging from beginner receive $50 in Atlas credits and free
to advanced levels. certification!

This concludes the material for this lesson. However, there are many more ways to
learn about MongoDB and non-relational databases, and they are all free! Check out
MongoDB’s University page to find free courses that go into more depth about
everything MongoDB and non-relational. For students and educators alike, MongoDB
for Academia is here to offer support in many forms. Check out our educator
resources and join the Educator Community. Students can receive $50 in Atlas credits
and free certification through the Github Student Developer Pack.