[Frank Fock]

Rachna,

For big tables, that do not user the RowStatus mechanism, the best
choice would be the to use MibComplexEntry class instead MibTable
and process request directly on the hardware or database.

For tables with row creation, you should subclass the MibTable class
and add only actually needed rows to the "virtual" table when the
MibTable::update(...) method gets called.

Best regards,
Frank

Rachna Bajaj wrote:

>  Hi Neel,
>   Thanks for the mail.I have already compiled my mib and generated the
> C++ code.By examples you mean the dynamic-table example?I looked into
> that.Actually I have one more question.My table is probably going to
> be very big with thousands of rows.So is there some way that instead
> of building this table in memory I can directly process the request by
> querying the sql server? Do you have a small sample which might show
> how to achieve this?My main doubt is how to process the request and
> when to add/update/delete the rows?
>
> Thanks,
> Rachna
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.agentpp.org/pipermail/agentpp/attachments/20020105/9d624f0a/attachment.htm