See the PPT, the inside of a data warehouse with data warehouse architecture, the architecture what basis points in the data layer, data buffer layer, application layer to the final data, it is logical layering or physical layer, such as a machine with the oracle database can realize,

CodePudding user response:

Oh, yes, with the schema

CodePudding user response:

Hierarchical data is not a one-sided statement, he is a data warehouse solution, the data architecture design logical structure of data, such as data buffer layer, basic data layer, the data warehouse architecture design means to enter the buffer layer, and then enter the basis data layer, the data layer is a logical, and physical and logical points according to these levels, physical will also build the layer in the process of implementation of data warehouse, such as data buffer layer may be a file, but the basic data layer is a physical table, through the ETL tool to realize from the buffer from the middle to process based data layer, data layer in the whole data warehouse implementation is in order to better for all levels of data management, each layer is the usefulness of each layer, such as: buffer layer, in order to reduce and source system coupling, avoid all of the data processing operations and direct source system, directly affect the normal operation of the original system, such as basic data layer, in the basis of data of a certain data layer will wash, standardization, integration, form the basis of available data, etc.,
Your question is the situation of data warehouse, and there are many concepts,

CodePudding user response:

General production environment, the data has granularity, data from the lower grain size is the basis of the data layer, through stored procedures or other etl tool processing to we want to collect data, that is, high granularity data from low to high level of granularity and the data granularity level conversion process is not a one-off, but gradually completed, it is a data buffer layer, you said in a production environment, buffer layer is very useful, because when the underlying data to the upper summary, we don't need to repeat some data processing, processing in the buffer layer at a time that can be invoked, the other upper table
For example: our company is an insurance company, using the oracle database, data is extracted from the core system through the interface to the interface table, processing into a low level of granularity of data table, then the interface table data through stored procedures processing (buffer layer) in the table into the buffer, buffer the data in the table is mild, comprehensive summary data buffer table data processing can be further to the upper level comprehensive summary table (fact sheets, also called the fact table), and finally the BI front desk can obtain data in a highly comprehensive summary table through interfaces, form the intuitive data reception,
Wish I could help you,

CodePudding user response:

Logical layering