Huinan wrote:
Hi,
I am tring to build an application that captures changes in Derby
database.For example, as a new query inserts data to some table comes, I
want to be able to get that change without using trigger for scalability
and loose coupling reasons.
The solution I'm considering is to use Derby's transaction log. I read
through its source code, espacially the package
org.apache.derby.impl.store.raw.log, trying to file clues about it's log
format and ways to instantiate log object. There is one class which
looks promising called Scan. But when trying to instantiate Scan, it
takes a LogToFile and a LogInstant. Both of them are not easily
instantiated. Does any one have a clue about how to access the log and
make sense of it? Or maybe some other ways to capture changes inside the
database?
As Mike also indicated, I am not sure how useful the log will be for
you. Since logging in Derby is "physiological", what you find in the
log is that some bytes of some record in some page has been changed.
You will not generally be able to determine the key of the record nor
which columns have changed.
--
Øystein Grøvlen, Senior Staff Engineer
Sun Microsystems, Database Technology Group
Trondheim, Norway
