I'm probably dense, but I don't see how it does.
rows.colnames has strings like 'foo.id' and 'a.name'. But you can't
index into a row with those strings: you can't say r['foo.id'] or
r['a.name']. You have to parse the column name and create code like
r['foo']['id'] and the problem I have with that is it seems like
improperly looking into the internals of the Row class. It seems like
the Row class should have some kind of method like r.col('foo.id')
that would "do the right thing". Or that r['a.name'] should just do
the right thing.
But, again, maybe I'm missing something.
On Oct 21, 11:24 am, Richard Vézina <ml.richard.vez...@gmail.com>
wrote:
> Maybe this thread could be of some help.
>
> http://groups.google.com/group/web2py/browse_thread/thread/e533b38bd5...
>
> <http://groups.google.com/group/web2py/browse_thread/thread/e533b38bd5...>
>
> On Thu, Oct 21, 2010 at 2:05 PM, BigBaaadBob <bigbaaad...@gmail.com> wrote:
> > Suppose you do a query like this:
>
> > rows = db((db.a.id==db.foo.xx) &
> > (db.b.id==db.foo.yy)).select(db.foo.id, db.foo.aaa, db.a.name,
> > db.b.name, db.foo.zzz, ....)
>
> > And you want to processes the results generically, say to produce JSON
> > for input to jqGrid. By generically, I mean not having to repeat the
> > "db.foo.aaa, db.a.name" stuff. You know, DRY.
>
> > For r in rows:
> > stuff...
>
> > What is the supported way of doing this?
>
> > Problems I see are:
>
> > * You can't say: r[0] because row is a dict with keys 'a', 'b', and
> > 'foo'.
> > * You can't say: r[rows.colnames[0]] because, well, the same reason
> > as the previous item.
>
> > I see code in the Rows class like this:
>
> > (t, f) = col.split('.')
> > if isinstance(record.get(t, None), (Row,dict)):
> > row.append(none_exception(record[t][f]))
>
> > But I wouldn't want to do that because it assumes the internal
> > structure of the Rows class...
>
>
