Naming and Sizing Conventions
Consistent usage of names and data types in an open database environment is important for user operability as well as administrative ease. Where users are free to run ad-hoc queries against one or multiple databases, the users should be able to expect consistent implementation of names and data types. Names and data types should be consistent both between and within databases.
Between databases, it is expected that certain core data elements will be named and sized the same between databases, regardless of the form (OLTP, ODS, Data Warehouse). Core data is defined as data which is shared between multiple databases and/or departments within the university. The Data Stewards will have the ultimate authority to determine what is considered to be core data. Listed below are initial examples of Core Data Elements, their Oracle Data Type and Size, and their Descriptions. This listing is not considered to be complete. It was arrived at after a brief review of the current major data structures. This listing may be considered a starting point for the Data Stewards in the development of a final listing.
|
Additional Naming and Sizing Standards
- For alphanumeric fields, the VARCHAR2 data type should always be used instead of the CHAR datatype.
- Oracle's maximum length for field names is 30 characters.
- All fields which contain a flag (such as "YES/NO" fields) must have the word, "FLAG," appended to their name.
- All fields which contain a code (such as STATE_CODE, CAMPUS_CODE, etc.) must have the word, "CODE" appended to their name.
- Fields which are not a flag or a code field should not have the associated word appended to their title.
- Field names, which are not listed in the above table, should be consistent within an application.
- Every effort should be made to make field names as descriptive as possible. This is especially important in an ad hoc database.