Subject Re: [ib-support] Re: Connect to DB from Procedure ??
Author Helen Borrie
At 02:44 PM 08-08-01 +0100, you wrote:
>At 12:46 08/08/2001, you wrote:
> >A-ha, have a look at the one-way Replication Module in the new version of
> >IB Objects. With this you'll be able to make a neat little NT service
> >module to take care of propagating these cross-database changes. It can
> >make use of idle CPU cycles too.
> >
> >Regards,
> >Helen
>Close but not quite right - it is not Replication but more like readings
>instructions from Database A and applying them to B,C,D,... That plus the
>changes to B,C,D.. (There are the same structure, A is different), must
>happen immediately.

It appears you haven't looked at how the IBO4 replication module works...the target structures can be completely different to the source structures, send data into multiple targets, aggregate data, whatever you want. You can set up a replication service that clears the replication cache every time there is spare CPU time, or pick up a timer event, or pick up an InterBase event, whatever you need, you can do. The cache stores pointers to changed rows, not the actual data, so the actual replication literally reads the changed rows and dispenses the selected data to the target.

On the target side, sets are inserted, updated or deleted as directed by a receiving procedure. It's all done with triggers and stored procs (which you write to suit yourself and can make considerably generic through the use of macro scripting...if you have to propagate the changes from A through to B, C and D you could set up a daisy-chain. You can set up inclusion and exclusion criteria on both sides of the wire.


>I know it is a bad idea to use UDFs for this but i cannot see another way.
>Damian Dowling
>IT Manager
>Pallas Foods Ltd
>Phone: + 353 69 20200
>Fax: + 353 69 20201
>Email: it-manager@...
>To unsubscribe from this group, send an email to:
>Your use of Yahoo! Groups is subject to

All for Open and Open for All
InterBase Developer Initiative ยท