{"id":1724,"date":"2020-07-30T07:00:00","date_gmt":"2020-07-30T04:00:00","guid":{"rendered":"https:\/\/www.dataplatform.gr\/?p=1724"},"modified":"2026-03-23T17:34:29","modified_gmt":"2026-03-23T14:34:29","slug":"ti-einai-ena-data-warehouse","status":"publish","type":"post","link":"https:\/\/www.dataplatform.gr\/en\/ti-einai-ena-data-warehouse\/","title":{"rendered":"What is a Data Warehouse?"},"content":{"rendered":"

In this article we will analyze what it is Data Warehouse<\/strong>, what you are made of and what is its use. <\/p>\n\n\n\n

In the age of information technology, the management of ever-increasing data is becoming more and more difficult. We want to make use of this data and at the same time not to delay our application.<\/p>\n\n\n\n

The Data Warehouse <\/strong>is a system you use for data analysis<\/strong>. There data is collected (ETL<\/strong>) in a database called staging <\/strong>from various sources such as transactional bases (OLTP<\/strong>) and Big Data <\/strong>and after some cleansing has been performed on the data (cleansing<\/strong> \/ data quality <\/strong>) these are transferred to its base Data Warehouse<\/strong> (OLAP<\/strong>) as small entities (Data Marts<\/strong>). From there, the Data Warehouse is connected to the Reporting Tools<\/strong> such as Power BI, Excel, QLik, Tableau, etc. so that the information reaches the end user. <\/p>\n\n\n

\n
\"\"
image is from Wikipedia<\/figcaption><\/figure>\n<\/div>\n\n\n

How is data distributed in a Data Warehouse?<\/h2>\n\n\n\n

The most popular data analysis technique is using Multidimensional cubes <\/strong>which are called and OLAP Cubes<\/strong> (Online analytical processing<\/a>) .<\/p>\n\n\n

\n
\"\"
image is from Wikipedia<\/figcaption><\/figure>\n<\/div>\n\n\n

There our data is divided into dimensions (dimensions<\/strong>) which can be time, product, geographical area, \u2026 and in fact <\/strong>where each cell contains a count (measure<\/strong>) which can be the number of sales made, the profit, the cost, etc.. Also the aggregations <\/strong>(eg average \/ total \/ bottom sales) <\/strong>are pre-calculated and stored when updating the cube with the new data through a process called process<\/strong>.<\/p>\n\n\n\n

How is data stored in a multidimensional cube?<\/h2>\n\n\n\n

The data of a multidimensional cube is stored in the form of either Star Schema<\/strong> either Snowflake Schema<\/strong> but before we see in detail what these two forms mean we should know the three terms below.<\/p>\n\n\n\n

Fact tables<\/h5>\n\n\n\n

To Fact table<\/strong> measurements are recorded (measures<\/strong>) of specific facts such as number of sales made, cost and profit. Also included are foreign keys<\/strong> which allow their connection with the dimension tables<\/strong>.<\/p>\n\n\n\n

To ensure the uniqueness of each record over time as changes may have been made to the source of the data, a unique number called a primary key is defined as Surrogate key<\/strong>. <\/p>\n\n\n\n

Dimension tables<\/h5>\n\n\n\n

Sta Dimension tables<\/strong> we have the dimensional data that may be common for measurements (measures<\/strong>) which we have in Fact tables<\/strong> such as time, employee, product and store.<\/p>\n\n\n\n

Its use Surrogate key<\/strong> to ensure the uniqueness of the documents it exists in them as well.<\/p>\n\n\n\n

Data Marts<\/h5>\n\n\n\n

Each separate entity of a subject such as for example finance or sales is called Data Mart<\/strong> and contains its own Facts Table along with Dimension Tables<\/p>\n\n\n\n

Star Schema<\/h3>\n\n\n\n

In a multidimensional Data Warehouse the simplest form of a Data Mart <\/strong>it is one Star Schema<\/strong>. Each Dimension Table is directly linked to the Fact Table through the Foreign Key.<\/p>\n\n\n\n

\"\"
image is from Wikipedia<\/figcaption><\/figure>\n\n\n\n

Snowflake Schema<\/h3>\n\n\n\n

That Snowflake Schema<\/strong> is a more advanced version of it Star Schema<\/strong>. Its difference is that Dimension tables are normalized into smaller sub-tables. Their use is recommended in cases where the speed of data recovery is more important than the recovery of detailed information.<\/p>\n\n\n\n

\"\"
image is from Wikipedia<\/figcaption><\/figure>\n\n\n\n

Advantages of using Data Warehouse:<\/h2>\n\n\n\n