Slow performance with long raw data

Question

Hi all,
we are running a 8i on linux (suse 6.2). everything is fine, but
when we try to select data from a table with two number and one
long raw fields the performance is real low.
data were exported from a 7.0.3 server running on a solaris box.
the new server is much more performant then the sun box, only
this one table doesnt work correct.
Any ideas what i did wrong? more details could be supplied if
needed.
Li
null

Answer 1

Caitanya wrote:
Blushadow,
I thought this is how a Cursor loop works:
..cut..
you need not fetch it again inside the loop? you fetch it once and the loop takes care of implicitly fetching one record at a time in the cursor.
Please correct me if I am wrong.Yes, typically a cursor loop starts the loop, then fetches and then tests for exit condition before processing.
The style the OP is using comes from the old BASIC style where we had REPEAT..UNTIL loops. The condition for these loops, due to their syntax, was specified on the UNTIL part e.g.
REPEAT
  .. code ..
UNTIL <condition>;This meant that the code would always be executed once and therefore, to prevent the code executing unecessarily you would have to code something like... (excuse my pseudo code)
IF <condition> = TRUE THEN
  REPEAT
    .. code ..
  UNTIL <condition>
END IF;So, in terms of what the OP is doing, they are fetching the first row outside the loop, using the condional check as the first thing inside the loop to prevent execution of the code if it's unnecessary, and then at the end of the loop fetching the next row before looping back around to the conditional check. Typically, you wouldn't do it that way, but it's fine and does work.