Database Reference
In-Depth Information
ii. In the condition of the SQL statements and SELECT sub-que
ries.
b. Explicit defi nition
i. INSERT column list.
ii. Columns to set in UPDATE statement.
2. Implicit usage
a. Implicit retrieval
i. Usage of * abbreviations to indicate whole table columns.
ii. Reference of master key column.
iii. Reference of detail foreign key column.
b. Implicit defi nition
i. Column of table used in DELETE statements.
ii. Columns not listed in the INSERT statement set to null value or
given a default value if it is available.
IMPACT ANALYSIS
Software impact analysis estimates what will be affected in software or related docu-
mentation if a proposed software change is made (Arnold & Bohner, 1996). Impact analysis
information can be used for planning changes, making changes, accommodating certain types
of software changes, and tracing through the effects of changes. Impact analysis provides
visibility into the potential effects of changes before the changes are implemented.
Although it is relatively easy to understand most of the database structures and
modules, understanding their combined effect or combined functionality is diffi cult. The
complex relationships between database objects make it diffi cult to anticipate and identify
the ripple effects of changes. Data dependencies, control dependencies, and component
dependencies make it diffi cult to generate tests to adequately retest the affected elements.
Our impact analysis technique is based on a reverse engineering approach designed to extract
the database components and their relationships. This information is used to automatically
identify the changes and the effects of those changes. In this section, we present phase 1
of our regression testing methodology, which includes modifi cation detection and impact
analysis. In this phase we localize the effects of change, identify all affected components,
and select a preliminary set of test cases that traverse modifi ed components.
Change Identifi cation
Change identifi cation is the fi rst step in change impact analysis. We differentiate be-
tween two types of changes in the database application environment:
(a)
Code Change
: This involves changes that can be made to the code of the database
modules. This is similar to any change made to any module written in any other lan-
guage. Addition, deletion, and modifi cation to particular statements inside a module
are examples of code change.
(b)
Database Component Change
: This change involves the changes that could be made
to the defi nition of the database components in general. It also includes the changes
that could be made to the defi nition of database modules.
