Re: Oracle Web Server row retrieval question

"Brad Worsfold" <bworsfold_at_vic.lgs.ca> wrote:

>> I want to write a stored proc for use with OWS that allows the
>> following behavior:
>>
 

>Oracle's WebServer Generator does this for the code that is generated.
>
>> I have a form for a user to specify the data to retrieve out of a
>> table. If the user specifies a query that returns a large number of
>> rows (more than 20 or 30), then I want to display the first 20 of the
>> rows, and have an anchor allowing the user to go to the next 20 rows,
>> or to the previous 20 rows.
>>
 

 Yep
>> The only way I can think of to do this is using a parameter to the
>> procedure that retrieves the rows that specifies the 'row number' to
>> start with, then use that to skip rows already seen, and also to use
>> it to quite the fetch loop.
>>

>> This solution seems flawed for a few reasons such as how much of a
>> HACK it is as well as the inability to guarantee that the result set
>> will be exactly the same on subsequent calls.
>>
 

>Definitely not! Webserver only allows for transaction processing. It will
>only know
>about the cursor while executing a single package. This gets tricky.
>> Is there a way to leave a cursor open on the server, or save the state
>> of it or something for the next call to the server?
>>

Here is the biggest problem I see with this: How do you know that the data has not changed between the diaplying of screens? How do you know that when you skip over the first 10 rows of a result set, they are the same 10 rows that were already displayed?

As far as it being a hack, I understand how OWS is doing it's thing, but it is a waste of resources to have to skip thorugh rows to get to a set of records.

>Hope this helps, good luck!
 

>Brad Worsfold
Received on Thu Jul 31 1997 - 00:00:00 CDT