However, I do have a question on this open improvement. The way I read the "improvement" is that the rows() method
will return all of the rows on the underlying tableset. To implement this I simply did a SELECT * FROM the table to
build the data. However, what happens if the user is trying to filter based on a where. Shouldn't the call really return all of the rows for the DataSet.

Please let me know how exactly this code should work. I can change the code to behave whatever way needed.

"... However, what happens if the user is trying to filter based on a where. Shouldn't the call really return all of the rows for the DataSet."

I think the answer to this is yes.

Here is some code:

def athletes = [
    [firstname: 'Paul', lastname:'Tergat', dateOfBirth:'1969-06-17'],
    [firstname: 'Khalid', lastname:'Khannouchi', dateOfBirth:'1971-12-22'],
    [firstname: 'Ronaldo', lastname:'da Costa', dateOfBirth:'1970-06-07'],
    [firstname: 'Paula', lastname:'Radcliffe', dateOfBirth:'1973-12-17']
]

athleteSet = db.dataSet('Athlete')
athletes.each { a -> athleteSet.add(a) }
youngsters = athleteSet.findAll{ it.dateOfBirth > '1970-1-1'}
def rows = []
youngsters.each { rows += it.lastname }
println rows.join(', ')           // => Khannouchi, da Costa, Radcliffe
println youngsters.rows().size()  // => 4

I would expect rows() for youngsters to return the three youngster rows not the entire original table (4 rows here).

public List rows() throws SQLException {
        String sql  = "SELECT * FROM " + table;
        return rows(sql);

    }

with

public List rows() throws SQLException {
        String sql  = getSql();
        return rows(sql);
    }

in DataSet?

Yes, your suggestion was almost what I did - just:

rows(getSql(), getParameters())

because sql is stored proc with params.