Data Warehouse is a subject-oriented information database, specially designed for reports and business analysis preparation to support decision-making within an organisation. It is built on the basis of database management systems and decision support systems. Data that enters the data store is usually read-only.
Data from an OLTP system is copied to a data warehouse in such a way that, during the construction of reports and OLAP-analysis, the resources of the transaction system are not used and its stability is not violated. There are two options for updating the data in the repository:
Complete data update in a repository. First old data is deleted and new data is then downloaded. The process occurs with a certain periodicity, while the data actuality may lag somewhat behind the OLTP-system;
Incremental update – only that data that has changed within the OLTP system is updated.
The principles of storage organisation
- Problem-objective orientation. Data is combined into categories and stored in accordance with the areas it describes, not with the applications it uses.
- Integration. Data is combined so as to meet all the requirements of an enterprise as a whole, and not the sole function of the business.
- Uncorrectable. Data in a warehouse is not created: that is, it comes from external sources and is not corrected or deleted.
- Dependence on time. Data in the repository is accurate and correct only if it is tied to a certain period or point in time.
Data sources can be:
- Traditional systems of operation registration
- Individual documents
- Extraction – moving information from data sources to a separate database, bringing them into a single format.
- Transformation – preparing information for storage in an optimum form for implementation of a request necessary for decision-making.
- Loading – putting data into a repository. This is performed atomically by adding new facts or adjusting existing ones.
- Analysis – OLAP, Data Mining, summary reports.
- Presentation of the results of analysis.
The difference between the Database and Data Warehouse
Speaking about the database, we analyse the singular data source, whereas with Data Warehouse have the figures collected from different sources and structured into one global storage. So, with Data Warehouse the information is wider and the architecture is much more complicated. However, it provides a better basis for decision-makers.