Saying Number Warehouse (1) - What is a Number Warehouse

Saying Number Warehouse (2) - Traditional Number Warehouse vs. Internet Number Warehouse

Saying Number Warehouse (3) - Number Warehouse Architecture

Saying Number Warehouse (4) - Metrics Dictionary

Saying Number Warehouse (5) - The Most Important Dimension of the Date Dimension

Saying Number Warehouse (6) - on the naming convention

Say number warehouse (7) - talk about data governance

Say number warehouse (8) - on the incremental

Say number warehouse (9) - upstream and downstream conventions

Say number warehouse (10) - task annotations

In recent years, the work and BI-related, but the number of real warehouse have a systematic understanding of or in the first company

The first time, when following IBM, IBM's business is to develop a new business, so that the business is not a business. The time, then followed the IBM consultant, to understand a lot of the story of the model design, for the beginning and enlightenment has a great help. The latter understanding of the number of warehouse, in fact, is the work of a little bit of practice and groping to get.

Previously, we said that the traditional number of warehouses and the Internet number of warehouses of the difference between the main reference to the different service user groups, here said that the traditional and the Internet are in fact relatively speaking, this time it, we say that the number of warehouses of the architecture.

Now that the number of warehouses, more and more data platforms or infrastructure on the ride, has been integrated into the entire infrastructure. Here it is, we do not say that the cooperation between the various components of Hadoop, we will briefly talk about the layered architecture of the number of warehouses.

Speaking of the number of warehouse modeling, we have to mention the classic 2 sets of theories:

The number of warehouse modeling or layering, in fact, is to better go to the organization, management, and maintenance of data, the actual development will be integrated into the 2 ways to use, of course, there are some other, like the Data Vault model, the Anchor model, for the time being has not yet been applied, not to say.

Dimensional modeling, generally referred to the star model, snowflake model, star model to do OLAP analysis is very convenient.

Simpler, direct ODS + DM can be, all the data will be synchronized, and then directly develop some of the application layer of the report, which is the simplest; when the DM layer of more content, want to reuse, it will be split into a public **** out of the layer into a 3-tier architecture, and recently read a book in Ali, "The Road to Big Data", which has a lot of warehouse-related content, very good. The dws layer is the bazaar layer, which is generally divided by topic and belongs to the category of dimensional modeling; the ads is the application layer, which is partial to the output of a variety of reports.

Based on this layered approach, it feels very convenient to use, well, first here.

Based on this layered approach, it feels easy to use.