Hi,

Have you tried chunking/paging the rowset? The javadoc says that
size() also returns the number of rows in the current page but maybe
this is not what you are looking for.

Kuassi http://db360.blogspot.com/2006/09/manipulating-tabular-data-using-jdbc_03.html

This has also been frustrating me. What I have discovered is that the
way the FilteredRowSet works is that once you have declared it and
applied your filter the actual filtered data has still not been
derived since working out the filtered set actually involves cycling
through all the data and applying your evaluate function defined by
your predicate. So for example lets say you ran a query and populated
your FRS and the applied a filter. I have written a basic predicate
that looks for exact matching values on columns called search

Search pred = new Search("gender",gender);
frs.populate(rs);
frs_ret.setFilter(pred);

At this point the Filter has not actually been applied so it does not
actually know how big it is. It is only applied when you call the
frs.next() function. And at this point it is only applied with one
evaluate function at a time i.e. not over the whole set

In this case it calls the evaluate function that checks to see if your
criteria applies and I assume it skips to the next one if it doesnt
apply. This is how it appears to work. So unfortunately it does look
like in order to work out the size of the filtered rowset you would
need to loop until the resultset is deemed to be finished. This
shoulnt post too much of a problem on a small resultset.

public int Size()
{
 cnt = 0;
 frs.beforeFirst();
 while (frs.next())
 {
   cnt ++;
 }
 return cnt;
 frs.beforeFirst();
}

This does seem rather inefficient but at some point the filter needs
to actually determine which rows apply and need to apply the tests
from your predicate. Perhaps there is a more elegant method. These are
my finding from running some tests on the FiltereRowSet and playing
with different predicates