![]() |
|||||||||||
| |||||||||||
|
Test Data ModelingThe main reasons for not copying the entire production data set into the test environment are size and complexity. Smaller test datasets are less resource intensive and make transaction tracing easier. SuperVisor simplifies the task of modeling the test dataset in two stages:
Determining Which Files are Required
A test dataset can be created without supplying a list of programs. In that case SuperVisor would assume that the entire dataset from the production library list would be involved in the test. However, if the purpose of the test dataset were to test only a single process or a narrow suite of programs SuperVisor can identify which files are required based upon a list of programs. A list of programs can be entered or selected using a SuperVisor analysis command. SuperVisor automatically expands the list of programs to include those programs called by those on the list. SuperVisor then performs an analysis to incorporate all of the files required by the programs to be tested. This includes all files required for referential integrity and related logical files. Filtering and Dependent RecordsTo reduce the number of records required to populate a test dataset,
SuperVisor provides two features, filtering and dependent copying. I Filtering is done by applying some logic which records are copied, such as comparing dates, customer numbers or item numbers. For example the order header file might be filtered by customer number. Order detail and history might copied using a dependent copy. The dependent copy would select only those detail records that matched those order numbers included in the order header file. Further an inventory transaction detail file might include only those items that were contained in the test order history file. Introducing a Conversion ProgramShould the system modification being tested require that certain files change in format, such as a field expansion, the test profile can specify the use of a user written conversion program. Stored in Test ProfileThese modeling techniques can be used each time the test dataset is repopulated. The specifications are stored in the Test Profile. So, they can be reused, modified and rerun. |
||||||||||

Copyright Software Management, Incorporated. For comments email: Software Management