> Reading here gives me some chance that it can be possible:
>
> http://www.firebirdsql.org/refdocs/langrefupd21-ddl-trigger.html#langrefupd21-alter-trigger-changecount
>
> What I want to do:
>
> I have two tables: MY_TABLE, MY_TABLE_HISTORY. In MY_TABLE_HISTORY I want to keep only one row, in this row I want to have an information such as:
>
> - last timestamp of changes
> - last change number(counter)
> - last row id which had been changed.
>
> I would like to update this row basing on the trigger created for MY_TABLE.
>
> However, there is one trick: if I update / delete/ insert many rows in one transaction I would like to update data in MY_TABLE_HISTORY only once.
>
> So for example: If I update 100 records in MY_TABLE in one transaction I am incrementing only once the last change number, but if I am updating 100 rows in 100 different transactions I am incrementing the last change number by 100.
>
>
> Is something like that possible? If not, do you have other ideas how to solve this problem?

Firebird doesn't have statement-levelm but only row-level triggers,
which means that a trigger fire only once per statement, independent on
how many records have been touched.

I haven't tested the following, but you might be able emulate that by
setting a flag in the trigger via RDB$SET_CONTEXT to know, hey my
trigger fired, then check that flag via RDB$GET_CONTEXT in an ON COMMIT
trigger to run some kind of post-process stuff with your desired
changes, e.g. processing logged rows from a temporary table etc.

Sounds a bit fishy though. ;-)



--
With regards,
Thomas Steinmaurer
http://www.upscene.com/