Using the Erwin Design database

Based on his years of practical experience, the author systematically expounds the thought, method and matters needing attention of using Erwin for database modeling, which has some practical value.

ERWin Data Modeler is a database modeling tool of CA company, which is widely used in the design of relational database. The author after many years of practice, feel the use of Erwin Design database, hands-on or relatively fast, but to use in the project, for different development environment and different projects, in the development of different stages of the use of Erwin, may adopt the best strategy is not the same.

Pre-use preparation

1. Learn the methodology of Erwin Support

ERWin support two methodologies, one is IE (Information Engineering), the other is IDEF1X, in the use of ERWin must understand the first, otherwise, will not even mark symbols are unclear. Here I have a brief talk about IDEF1X (details are presented in Erwin's online documentation). IDEF1X provides a normative structure for the data model and is a semantic modeling technique, which mainly describes the entities, links, and attributes. At the same time, as an industrial specification, IDEF1X also emphasizes the methods needed to develop the above model. In this way, the combination of standardized markup language and related auxiliary methodologies can fully guarantee the balance of efficiency and effectiveness of the design.

2. Learning Erwin

Mastering the methodology of Erwin Support does not mean mastering the Erwin, the methodology only solves the logical model, but Erwin also supports the physical model, as well as the interface and operation problems. Because in the process of building the database, need for the use of the physical database has a lot of understanding, so we must understand the difference between the IDEF1X and the target relational database, this kind of difference may have little relation to the microcomputer platform, small amount of data application, but for large database, there are many physical parameters, restrictions, etc. should be understood.

4. Determine the naming rules for database tables and fields

Determine the database table, field naming rules, seemingly easy, in fact, there are many aspects involved, and the initial stage once not handled well, later to change the difficulty is greater. The author believes that the naming should consider the following factors:

If the newly developed system is a subsystem of a large system, then should consider the original large system database, field naming rules, even if such a rule has problems, but also to achieve consensus on the basis of improvement.

Consider the limitations of the development and Operation Tools, as well as the production system requirements.

Where possible, the Hungarian nomenclature should be considered. For application systems, there is often a classification of data, which is beneficial if it can be embodied in database table names and field names.

The field name remains the only one that avoids some careless use of database fields.

5. Classification of database tables

The classification of database tables makes the database clearer and easier to manage. According to the author's experience, for each type of database table, if allowed, according to the Hungarian Nomenclature rules of a feature tag, can be a prefix or suffix.

Establishing a logical model of the database

As a modeling tool, Erwin introduces concepts and tools that often run through the logical and physical models. However, if these aspects are not handled well in the logical model, the phase of the physical model is often done, and there is no way.

1. Use good domain

Domain concept is a bit like the data type of the attribute, the author's experience is that if you do not intend to use domain, do not add any domain, with the default value provided by Erwin; If you intend to use domain, you should create a domain for each type of data equivalent, and modify the data type to change only the content in domain. Overall, using domain may add some work, but you can establish a definition tree for all attribute data types.

2. Use good definition

Definition and domain are not an operable entity, but a label in every entity and attribute in domain. In the process of establishment and modification of domain, entity and attribute, it is an effective method that can get the description of text data structure at any time by correctly maintaining definition.