Forgive me if I don't make sense, because I didn't understand
everything you said, but...
Are you actually experiencing performance issues here? I find it
extremely hard to believe that you are. If something is slow, I'd look
elsewhere.
Second, your query should be named (query.setName()), or the caching
won't be quite what you have in mind.
Cris
On 1/18/06, emre.yilma..r.net <emre.yilmaz@tr.net> wrote:
> Hi,
>
> Actually I want to find different personel cards from PersonelCard table with
> its cardNo entity(its actualy unique) for every second. I think if I hold the
> PersonelCard table in the cache, this can be increased the performanse and when
> I want to find the personel card with its cardNo entity, I use the caches data
> for searching and finding the PersonelCard row and its relational objects. So I
> think I haven't to do query for finding the each PersonelCard row from the
> sql.
> In userGuide for caches query advice the following:
> performQuery(java.lang.String queryName, boolean refresh)
> if refresh is false.
> but i don't use the a named query stored in one of the DataMaps.Instead i use
> below SelectQuery and i quess
> query.setRefreshingObjects(false) indicates that a cached result should be used
> if available. Is it correct?
> Did you mean setting pagesize for "is CARD_NO indexed". How do I query with
> cached data ?
>
> > > Code:
> > > Expression qualifier = ExpressionFactory.matchExp(
> > > PersonelCard.CARD_NO_PROPERTY, cardNo);
> > > SelectQuery query = new SelectQuery(PersonelCard.class,qualifier);
> > > query.setDistinct(false);
> > > query.setCachePolicy(GenericSelectQuery.SHARED_CACHE);
> > > query.setRefreshingObjects(false);
> > > query.setFetchingDataRows(true);
> > > List cards = context.performQuery(query);
> > >
> >
> >
> > I'm with Tore... in this case, DataRows aren't going to do you any
> > good. DataRows don't make data queries faster, just object
> > construction... and since you're only constructing 1 object....
> >
> > On the other hand, is CARD_NO indexed? A repeated table scan is
> > expensive. Indexes rule. If its not, distinct makes it even worse - I
> > would avoid using distinct here, particularly since you're not doing a
> > join. Just ignore the second, third, etc results if you only want one.
> >
> > Finally, why are you running this every second if its cached???
> >
> > Cris
> >
> >
>
>
This archive was generated by hypermail 2.0.0 : Wed Jan 18 2006 - 09:56:29 EST