Manual replication using snapshot logs

Hello,

is it recommended to use snapshot logs as logs of change when copying data manually to other server?

I have a couple of tables that need to be replicated to a remote Informix server. I use Perl and DBI to connect to both databases and copy the data. In order to be able to copy the data more or less in real time (each 5 minutes is enough) and to minimize the amount of data transferred, I want to have a list of changes, so that I only do the minimum necessary amount of work.

One possibility is to create triggers for each row on those tables, copy the primary keys or rowids to some "log" table in these triggers and then have the Perl cron job fetch the changed data according to these logs.

Another one is to use snapshot logs that would do the exact same thing for me -- fill the mlog$_table table with changes. That seems a little bit neater than the approach with triggers. But will Oracle be happy about this? Should I use dbms_snapshot.purge_log to clear the log or just do delete? How would this mix with Oracle's internal notion of snapshots if I touch the logs from my external process?

Unfortunatelly I couldn't find any info about using snaphshot log for anything else than snapshots.

Comments or points to documentation would be greatly appreciated.

Yours,

-- 


------------------------------------------------------------------------

 Honza Pazdziora | adelton@fi.muni.cz | http://www.fi.muni.cz/~adelton/
 .project: Perl, mod_perl, DBI, Oracle, auth. WWW servers, XML/XSL, ...
Petition for a Software Patent Free Europe http://petition.eurolinux.org


------------------------------------------------------------------------