Re: Does a trigger and the event that trigger follow separate processign paths?

tried an raise_application_error exception yet process was able to do insert. Is their a more appropriate exception to handle this? thnaks for inf0

In article <d0Y73.27229$Fz2.4026_at_news.rdc1.az.home.com>, "Kevin P. Fleming" <kfleming_at_access-laserpress.com> wrote:
>If you have a BEFORE INSERT trigger that raises an exception, the INSERT
>will be aborted. It's up to the application as to how it handles that, but
>the rows with bad entries definitely won't make it into the table.
>
>ramdan <ramdan_at_mailexcite.com> wrote in message
>news:l_X73.6503$wk2.151551_at_newscene.newscene.com...
>> We have a process that inserts rows into a table but does not do good
>error
>> checking. We can't change the process so we need to do error checking
>after
>> the fact. We will reject records that have a specific value in a non
>> key non unique varchar field. The process is reading a text file and
>> inserting it into a table
>>
>> three oprions have been proposed and any input would appreciated
>>
>> 1- have a trigger that does a before insert looking for the specific value
>in
>> the incoming data and stopping the insert or causing it to fail. This
>getts
>> at the subject line if the trigger tries to stop insert what happens? do
>they
>> follow spearate paths or does the insert wait for trigger to finish
>>
>> 2 - have a trigger that does a delete after to get rid of bad data. this
>> however consumes resources and lets bad data into table even if for a
>short
>> time
>>
>> 3 - maybe somekind of constriant that prohibits the bad value from bieng
>> inserted?
>>
>> any ideas greatly appreciated
>
Received on Thu Jun 10 1999 - 20:43:09 CDT