Database Models – Part 2

April 8, 2019March 26, 2021 Kavinda

You can find the Part 1 of article on Database Model here

Key-Value Database Model

Key value database is a data storage paradigm where data is stored, retrieved and managed in associative arrays. These are now commonly known as Dictionary or Hash Table. As the name suggest it consists of Key and Value pairs. Here keys are unique and values can contain data of various lengths and types. He user can access the value using the key. There are no predefined schema for the values so it may contain different field for every record.

Advantages of Key Value Model

Faster read and write operations due to simple data format
Value can be anything including JSON

Disadvantages of Key Value Database Model

Optimized only for data with single key and value.
Not optimized for lookup

Column-Based Database Model

In a column based database data is stored in columns rather than in rows. This type of databases are mainly used in large data warehouses. The reason for this is that when doing analytical operations accessing only the required columns is much faster than accessing each column and then extracting few columns you need.

Advantages of Column Based Database Model

Compression
- Highly efficient data compression
Aggregation queries
- As the column based databases can query directly to columns they perform well in queries such as calculating sum, calculating average etc.
Scalability
- Column databases are highly scalable
Fast to load and query
- These databases can be loaded extremely fast.

Disadvantages of Column Based Database Model

High Insertion Time
High update time
Low performance for queries with JOINS

Document Database Model

Document oriented database model is designed for storing, retrieving and managing document oriented semi structured information. This can be described as a subclass of key value database model. Difference is that in Key Value database model data in the value is inherently opaque to the system but in document database model, system relies on meta data in the value to further optimizations. Generally the data is encoded using standard encoding techniques such as XML, YAML, JSON etc.

There is no fixed schema for the structure of the document. So it is easy to insert new field later.

Advantages of Document oriented database model

Documents are independent
Objects can be directly written as documents
Can easily store unstructured data
Have strong indexing feature

Graph Database Model

In graph database model graph structures such as nodes and edges are used to represent and store data. Graph data model can be referred as being “whiteboard-friendly”. Which means we can directly build the database using arbitrary data and their connections without formatting and structuring to fit a normalized model.

Advantages of Graph data model

Object oriented thinking
Very high performance for querying related data.
Can solve problems which are impractical for relational query.
Can do real time updates on big data simultaneously querying the data.
Flexible
Can do group by aggregate queries which are not possible in relational databases.
Can combine multiple dimensions to manage big data.