Relational VS Non-relational DB

Relational Database

• Store data into table

  Row represents entity and Column represents attributes.

• Each table has a primary key

  Data in the table is linked to other tables via relations consisting of primary key - foreign key pairs.

• Use SQL(Structured Query Language) to query.

Example:

• Table: Weather
• Columns: Days of the week
• Rows: Time of Day
• Data: Degrees

Non-relational Database

• Built for unstructured data:
  1. Key-value stores
  2. Document
  3. Graph
  4. Wide-column stores

  Store information into rows and columns, but rows have different values in different columns

Example: FaceBook Messenger

Comparison

1. Non-relational databases offer high availability - always available for read and write even when faced with network failure.

   Relational databases focus on high consistency.

2. Relational databases are usually developed over a single server or node. Scaling is slower and more expensive.

   Non-relational databases are distributed by design. Scaling is faster and cheaper.

3. Relational databases are best for structured data by table model.

   Non-relational databases are best for different data structure.

4. Non-relational databases are designed to handle big amounts of data which may break a relational database.