Data warehouse concept
1. Concept Design Model
The work is to be done:
(1) define system boundary
What are the types of decisions do?
Policy makers are interested in what is the problem?
What kind of information on these issues need?
To get this information needs to include the existing database what data?
(2) identify the main theme and its content:
It is based on the theme of business, not the technology itself. If a business can be divided several independent modules according to a certain scale, each module is a theme that domain. This is the idea of divide and rule
Meaning customers can simply give you the theme of the collection is interpreted as the theme to the customer the customer number as the primary key of the fact table;
which contains information on all aspects of the customer;
The data warehouse is the original data in the database integration and reorganization of the collection, see the original data in the "What" and "How the organization", "how they are distributed," and so on, to consider how to build a data warehouse
Data warehouse steps
1. Determine the theme
Theme and business combination, before designing the data warehouse must fully understand the business aspect of what demand there is, accordingly determine the theme
2. Identify metrics
Determine the theme, considering technical indicators to be analyzed
3. Determine data granularity
Taking into account the different levels of aggregated measure, we will use the "minimum granularity principle", is about to measure the particle size is set to the minimum
4. Determine the dimensions
The design of the main keys of each dimension, hierarchy, hierarchy, to reduce redundant
5. Create Fact Table
And each of the key dimensions agent present in the fact table measure, there should not descriptive information, i.e. matching "lanky principle", i.e. the fact table requires as many number of pieces of data (the minimum size), and descriptive information as little as possible.