No code duplication in loops

[Ursego]

Link to this posting

by Ursego » 21 Feb 2013, 10:48

Don't populate the variable, used for looping control, twice (before and inside the loop body). Populate it only once, inside the loop.

You can ask - why? We see that kind of duplication in most programming books! The answer is simple: we have to avoid any sort of code duplication, so there is no reason to write twice something which can be written only once to achieve the same goal.

*** BAD code: ***

PB:

Code: Select all

ll_row = dw_cust.GetNextModified(ll_row, Primary!)
do while ll_row > 0
   [do something with the modified row]
   ll_row = dw_cust.GetNextModified(ll_row, Primary!) -- deja vu!
loop

PL/SQL:

Code: Select all

FETCH emp_cursor INTO v_row;
WHILE emp_cursor%FOUND LOOP
   [do something]
   FETCH emp_cursor INTO v_row; -- deja vu!
END LOOP;

An eternal loop with 'break' ('exit when' in PL/SQL) is a very elegant solution in such a situation:

*** GOOD code: ***

PB:

Code: Select all

do while true
   ll_row = dw_cust.GetNextModified(ll_row, Primary!)
   if ll_row < 1 then exit
   [do something with the modified row]
loop

PL/SQL:

Code: Select all

LOOP
   FETCH emp_cursor INTO v_row;
   EXIT WHEN emp_cursor%NOTFOUND; -- or "IF emp_cursor%NOTFOUND THEN BREAK;"
   [do something]
END LOOP;