 |
|
Pseudocode for the
Parse_String() Function
Oracle Tips by
Burleson
|
Pseudocode
It’s often useful when designing a module to generate pseudocode
that outlines the logical steps that the module must take. Listing
5.21 shows a very simple bit of pseudocode for the Parse_String()
function.
Listing 5.21 Pseudocode for the Parse_String()
function.
store the string to be parsed in a local variable;
enter a loop
get the location of the caret inside the string;
if there is no delimiter in the string then
return the array to the calling procedure;
store the first section of the string in the array;
chop off the first section of the string;
increment the index variable for the array;
if the VALUE_ERROR exception occurs then
log an error using the system log function;
This approach is especially useful with procedures, but functions
often handle some very tough problems, too. Breaking the functionality
of the module down into steps is the real work when writing code. Once
the logic for the function has been thoroughly defined, the code can
be written without too many problems.
Comments
Used thoughtfully, comments are an excellent tool for documenting
code. The best comments describe why code works the way it does
instead of describing how it works. Consider the following two
sample comments:
Comment A:
--
-- If the employee has been late for work less than 1% of the time,
-- grant the employee a 0.5% raise.
--
IF (nOntimePercent > 99.0) THEN
nRaiseAmount := nRaiseAmount + 0.005;
END IF;
Comment B:
--
-- If nOntimePercent > 99 add .005 to nRaiseAmount.
--
IF (nOntimePercent > 99.0) THEN
nRaiseAmount := nRaiseAmount + 0.005;
END IF;
Comment A explains why the code is written in a particular way.
Comment B paraphrases the code but doesn’t explain the business rules
behind the code. There is nothing in comment B that will help you
understand what business rules the code satisfies. The sample PL/SQL
coding standard in Appendix D includes some guidelines about the
content and location of comments.
Identifiers
Consider this example:
IF (x > 99.0) THEN
y := y + 0.05;
END IF;
It might take you a moment to recognize this block of code as the
code from the previous example, only with different variable names.
While it’s relatively easy to keep track of x and y in
this example, when repeated several times in 200 lines of code, x
and y will become painfully and hair-wrenchingly obscure.
The only time variable names like x and y are
potentially meaningful is when referencing a loop control variable or
the index of a PL/SQL table. Even then, it’s better to give variables
names related to their functions. Using meaningful identifiers is one
of the easiest ways to document code.
This is an excerpt from the book "High Performance Oracle
Database Automation" by Jonathan Ingram and Donald K.
Burleson, Series Editor. |
