Explore topic-wise InterviewSolutions in .

NoSQL database is a type of database that does not require a fixed schema. It is non-relational and non-relational and usually avoids joins in preference of schema variation for achieving better performance. It is also very easy to scale based on the application’s NEEDS. NoSQL is used for storing huge data in a distributed manner. NoSQL DATABASES are of four types:

• Column-based - These organize data in columns instead of rows. The rows in these databases have many columns associated with them. These are useful in fraud detection systems, catalogues, recommendation engines etc. Example: Cassandra, Hypertable, Hbase, Amazon DynamoDB etc
• Graph-based - These organize data into nodes and edges. The edges show connections between different nodes. It is multi-relational. These are mostly used in social networks, LOGISTIC domain etc. Example: Neo4J, OrientDB, Infinite Graph, FlockDB etc
• Document-based - These types store semi-structured data in the form of documents. These are similar to key-value type databases and the difference is that the values in this type store values in the form of JSON, BSON, XML etc. Example: MongoDB, RavenDB, Terrastore etc
• Key-value - SIMPLE NoSQL data model consisting of unique keys and values linked with it. It performs efficiently and is highly scalable. Redis and Memcached are some examples of this database type and is mostly used in caching applications, session MANAGEMENT.

Cardinality REPRESENTS the level of uniqueness of DATA values present in a column. The high value of cardinality TELLS us that the column has a large percentage of UNIQUE values. Low cardinality indicates that the data is repeated (duplicated) in the column.

A RELATIONAL database is a DIGITAL database system that is based on a relational data model and is useful for storing data in tables. Oracle, Teradata, MySQL, PostgreSQL, SQLite etc are types of relational databases. The data is TRANSFORMED into rows and columns which is STANDARDIZED and can be queried with SQL. RDBMS helps to add/ alter tables and column data into the database systematically by maintaining data integrity.

Metadata is the INFORMATION about data that tells what data KIND is stored in the system, what is the PURPOSE and to whom it is intended. There are different types of metadata depending on the purpose, they are:

• Technical Metadata – This type DEFINES names of the database system and tables, size of tables, values, attributes, data types etc. It also INCLUDES constraint information like primary keys, foreign keys, indexes etc.
• Business Metadata – This data is specific to business and defines the data ownership, business regulations and rules, policies etc.
• Descriptive Metadata – This type of data gives information about folder, file, image, book or video. Information can be title, date, size, author, published on etc.
• Operational Metadata – This type includes information related to any business operation and is used by managers or executives for performing any task.

OLAP STANDS for On-Line Analytical PROCESSING and it is a type of technology that authorizes managers, executives and analysts for gaining insights in a faster, secure, consistent and interactive manner. OLAP is USED in Intelligent solutions involving Planning, Budgeting, Analysis, Forecasting, Simulation Models etc. OLAP helps clients to perform analysis considering MULTIPLE dimensions and helps in providing insights that help in better DECISION making.

OLTP stands for Online TRANSACTION Processing which is an operational system supporting transaction-oriented applications in 3-tier ARCHITECTURE. It focuses on QUERY processing, DATA integrity maintenance, transaction EFFECTIVENESS per second in multi-access environments.
Some of the examples of OLTP systems are:

• Sending text messages.
• Online banking.
• Online airline ticket booking.
• Order entry.
• Adding items to the shopping cart.

Database MANAGEMENT System (DBMS) is SOFTWARE CONSISTING of a group of PROGRAMS that manipulates the database for storing and retrieving the user DATA.

DATA sparsity refers to the number of empty cells in the database. It represents how much data is available for a particular dimension in a data model. Inadequate information leads to large space CONSUMPTION for SAVING aggregations.

Granularity represents the information level stored in a database table. It can be high or low level with tables that CONTAIN transaction-level or fact table RESPECTIVELY. Granularity is ALSO a measure of smallest set of DATA which can be INDEPENDENTLY processed by any component/task/application.

DIMENSIONS represent the SET of unique values that are useful for IDENTIFYING and CATEGORIZING the data from the data WAREHOUSE.

A data model is an information schema that sorts and normalizes different information COMPONENTS and establishes relations between them. These models subsequently become tables in the database which can be retrieved and processed based on the organization’s requirements. There are three types of data models, they are:

• Conceptual data model:  This TYPE of model focuses on high-level, ABSTRACT and users view of data representation. It defines what the system contains and is created by data architects and business stakeholders. The main purpose of this model is data organization and defining the scope and business rules. Following are the characteristics of the conceptual data model:
  • This provides an overall overview of data structure in terms of business context.
  • Contains some concepts that are not required for database implementation but are required for a business for understanding the need for features in the applications.
  • Represents data needed for performing business processes and operations.
• Physical data model: In this model, we define the database schema and how the data is stored in the database PHYSICALLY. Following are the characteristics of the physical data model:
  • It describes the data required for the application. It can be integrated with different physical models depending on the scope of the application.
  • This defines relationships between different tables which addresses cardinality and nullability of relationships.
  • This model is developed for specific DBMS versions, data storage, locations or technologies used in the application.
  • This model also consists of exact data types, constraints, lengths and default values.
  • It also defines primary and foreign key references, indexes, views, profiles and authorizations.
• Logical data models: This data model lies between the conceptual and physical model and describes the logical representation of data. Following are the characteristics of the model:
  • It describes the data needs of a project and can also be integrated with other logical models depending on the scope of the application which in turn guides the physical model.
  • This data model is designed and developed independently from DBMS.
  • It also consists of attributes like data types and exact length and precisions.
  • It contains normalization processes of the model and is applied till it ACHIEVES 3NF.

Data modelling helps professionals from different domains like software development, business ANALYTICS, data science, etc to view the data, understand the relationship between the data objects in a data warehouse or a database. Following are the main benefits of this process:

• Reduce errors during the software development
• IMPROVE the performance of the application and the database.
• Ease the process of mapping data between the different processes in an organization.
• Helps to improve communication between the business intelligence and the developer teams.
• ENHANCES the speed of database design at CONCEPTUAL, logical and physical levels.
• Proper data modelling process helps to reduce the cost of maintenance when the system grows at a large scale.
• It also supports the reusability of the data models when the requirements are repetitive
• Improves the quality of data and enhances the robustness of the application.

Data Modelling is defined as the PROCESS of creating a model that REPRESENTS the data and the relationship between different data to store it in a database. It can also be called database modelling and this skill is useful across all domains like data engineering, data science, software development etc to prepare, analyse and process the data by CONTINUOUSLY REORGANIZING, restructuring and optimizing it for fitting the needs of the company/organization.