Shulou(Shulou.com)06/02 Report--

This article shows you how to judge the quality of a data model, the content is concise and easy to understand, it can definitely brighten your eyes. I hope you can get something through the detailed introduction of this article.

High cohesion and low integration

The data with similar business or related data with the same granularity is designed as a logical or physical model: the data with high probability and query at the same time are put together, and the data with low probability accessed at the same time are stored separately.

Separation of core model and extended model

Establish the core model and the extended model system, the core model includes the Yu segment to support the commonly used core business, and the fields included in the extended model support the needs of personalization or a small number of applications, so that the Yu segment of the extended model cannot excessively invade the core model so as not to destroy the architectural simplicity and maintainability of the core model.

L common processing logic sinking and singularity

The more the underlying common processing logic should be encapsulated and implemented at the bottom of the data scheduling dependency, do not expose the common processing logic to the application layer, and do not let the common logic exist in multiple places at the same time.

L balance cost and performance

Appropriate data redundancy can be exchanged for query and refresh performance, not excessive redundancy and data replication.

L data can be rolled back

The result of the rerun task remains unchanged without changing the processing logic and without modifying the code.

L consistency

Field naming and definition must be consistent

L the name is clear and understandable

Table names should be clear and consistent, and table names should be easy to understand.

So, how to judge whether a model is good or bad?

The general principle should be

L improved demand response speed

L query speed is faster

L query cost reduction

L improvement of user satisfaction

1. Integrity of the model layer

For a better model, users can get all the desired data directly from the model. If the proportion of tables directly referencing the ods layer in the dws,ads,dm layer is too large, that is, the cross-layer citation rate is too high, then the model is not optimal and can continue to be optimized.

two。 Reuse degree

The number of tables directly output downstream of dw,dws

3. Degree of standardization

Tables need to be associated with topic fields and need to be layered

Table naming conforms to the specification (clear and consistent, and the table name should be easy for users to understand)

Field naming depends on the root

4. Data can be rolled back

In the case of rerunning the data, the data result remains the same.

5. Separation of core model and extended model

Establish the core model and the extended model system, the core model includes the Yu segment to support the commonly used core business, and the fields included in the extended model support the needs of personalization or a small number of applications, so that the Yu segment of the extended model cannot excessively invade the core model so as not to destroy the architectural simplicity and maintainability of the core model.

The above is how to judge whether a data model is good or bad. Have you learned any knowledge or skills? If you want to learn more skills or enrich your knowledge reserve, you are welcome to follow the industry information channel.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.