NoSQL Database
# Introduction to NoSQL
NoSQL is a technology that represents a class of products that does not follow RDBMS
principles and are often related to storage and retrieval of massive volumes of data.
NoSQL databases are structured in one of the following ways: key-value pairs, document-
oriented database, graph database, or column-oriented database.
Why NoSQL
NoSQL databases support large volumes of structured, unstructured, and semi-structured
data.
As NoSQL databases are schemaless, integrating huge data from different sources becomes
very easy for developers, thus, making NoSQL databases suitable for big data storage
demands, which require different data types to be brought into one shell.
CAP Theorem
CAP is the acronym for consistency, availability, and partition tolerance formulated by Eric
Brewer.
According to Brewer, a database cannot exhibit more than two of the three properties of the
CAP theorem.
ACID
ACID is the acronym for a set of properties related to database transactions. The properties
are atomicity (A), consistency (C), isolation (I), and durability (D).
Relational database management systems exhibit ACID properties.
BASE
BASE is the acronym for a set of properties related to database design based on the CAP
theorem.
NoSQL databases exhibit the BASE properties.
Schemaless Databases
Schemaless databases are those that do not require any rigid schema to store the data. They
can store data in any format, be it structured or unstructured.
# NoSQL (Not Only SQL)
NoSQL database is a non-relational database designed to store and retrieve semistructured
and unstructured data.
Unlike the traditional database systems, which organize the data in tables,
the NoSQL database organizes data in key/value pairs, or tuples. As the name suggests, a
NoSQL database supports not only SQL but also other query languages, namely, HQL to
query structured data, XQuery to query XML files, SPARQL to query RDF data, and so
1
forth. The most popular NoSQL database implementations are Cassandra, SimpleDB, and
Google Bigtable.
# NoSQL vs. RDBMS
RDBMS are schema-based database systems. It takes a significant amount of time to
define a schema, but the response time to the query is faster.
NoSQL (Not Only SQL) databases have the capability to store the data in HDFS as it arrives
and later a schema can be defined using Hive to query the data from the database.
2
# Features of NoSQL Databases
Schemaless
Horizontal scalability
Distributed computing
Lower cost
Non-Relational
Handles large a volume of data
# Types of NoSQL Technologies
1) Key-value store database
Examples: Amazon DynamoDB, Microsoft Azure Table Storage
2) Column-store database
3
Example: Apache Cassandra
3) Document database
Example: CouchDB
4) Graph database
4
Example: Neo4J, Cypher Query Language (CQL)
# NoSQL Operations
Create database
Drop Database
Create collection
Drop collection
Insert Document
Update Document
Delete Document
Query Document
# Migrating from RDBMS to NoSQL
- Drawbacks of traditional relational database are:
● Entire schema should be known upfront
● Rigid structure where the properties of every record should be the same
● Scalability is expensive
● Fixed schema makes it difficult to adjust to needs of the applications
● Altering schema is expensive.
- Below are the advantages of NoSQL, which led to the migration from RDBMS to NoSQL:
● Open-source and distributed
● High scalability
● Handles structured, unstructured, and semi-structured data
● Flexible schema
● No complex relationships