Re: polling, triggers, java, Oracle, C++

From: <nonick.hughespam_at_icllll.com.invalid>
Date: 2000/06/20
Message-ID: <8int6q$24sf$1_at_news.icl.se>#1/1


Markus Laun wrote in message <394F5CBB.7565689A_at_cdsoft.de>...
>I also think your answer was rude and not very constructive!
>

I'll go along with that. A few more constructive comments below.

>Sybrand Bakker wrote:
>>
>> Comments embedded
>> <argosy22_at_my-deja.com> wrote in message
 news:8ilg7j$n44$1_at_nnrp1.deja.com...
>> > HI all,
>> >
>> >
>> > There are java processes that poll the database.
>> > Sometimes, they use the API of another product, which
>> > also polls the database. There are also c++ processes
>> > running on NT clients, which poll the database.
>> >
>> > The idea is that customers will use the web interface,
>> > and insert a record into the database. The polling
>> > then checks for new records.
>> >
>> > The problem is that we are running the database in archive
>> > mode.
>> So every select is not only logged
>> , the logs are
>> > later archived when they fill up.
>>
>> This is nonsense, selects are not logged
>>

Clearly the select *is* being logged, perhaps one of the application layers is declaring read/write transactions (for which AIUI the Start and commit/rollback would be logged regardless of any updates inbetween).

>> >
>> > My initial thoughts (coming from an Oracle background) is
>> > to use an Oracle trigger. On INSERT ....
>> >
>> > One question was, how to call an executable java file (or even a
>> > ProC program) on the Unix side from the PLSQL in the trigger?
>> > Is it possible in 734? I don't think that it is. How about with 8?
>> >

Oracle 8i supports java triggers, I've never used it myself so I can't comment on the performance but you store them within the database in the same way as you would a PL/SQL trigger. I think you would need to beg/borrow a copy of the Java Stored Procedures manual for 8i.

--
Nic Hughes

The views of the author may not necessarily reflect those of the Company
Received on Tue Jun 20 2000 - 00:00:00 CEST

Original text of this message