Mailing-List: contact users-help@jackrabbit.apache.org; run by ezmlm
Precedence: bulk
Reply-To: users@jackrabbit.apache.org
Received-SPF: pass (nike.apache.org: local policy)
Content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
Subject: AW: AW: How do I get the total size of nodes/rows from a QueryResult
Date: Wed, 26 Oct 2011 09:22:09 +0200
Message-ID: <DD48F4D5B13C134A81F9FB51AD43DE21027F6F89@email.innoimm.local>
In-Reply-To: <CACCF310.8F732%aklimets@adobe.com>
Thread-Topic: AW: How do I get the total size of nodes/rows from a QueryResult
Thread-Index: AcyTW+ak2uNsw70aRq6/NlEj5GkFewAUte/w
References: <DD48F4D5B13C134A81F9FB51AD43DE21027F6F16@email.innoimm.local>
 <CACCF310.8F732%aklimets@adobe.com>
From: <Sascha.Theves@bosch-si.com>
To: <users@jackrabbit.apache.org>

> You can't be using limit & offset with Jackrabbit 1.x and Xpath...

Hm but I really do that and it worked:

   	((QueryImpl) query).setLimit(select.getLimit());
   	((QueryImpl) query).setOffset(select.getOffset());
   	QueryResultImpl queryResult =3D (QueryResultImpl) =
Executor.execute(query);
	queryResult.getTotalSize();

> I guess it's only that getSize() returns -1 for you now in most cases =
(if
> you come from Jackrabbit 1.x), as that is an optimization in the =
engine.
> To force a result, use the orderby @jcr:score trick mentioned by Ard.

Thanks for the tip and that makes 50% of my test cases to run without =
error but=20
the other 50% do not work. I stripped the error down and found out that =
each query that uses a "contains" condition fails now.
But the test fail not deterministic and I think that is because I add =
nodes with XML content (binary) in my test cases and immediately after =
adding the node I am searching for the content of the node. I think that =
the XMLTextExtractor runs in a paralles Thread and the computation is =
not finished when I try to query jackrabbit. I could now implement a =
delay in my tests but the best solution would be to wait until the =
XMLTextExtractor job is finished. Is there a possibility to do that or =
am I completely wrong with my suggestion?

Thanks for your help!

-----Urspr=FCngliche Nachricht-----
Von: Alexander Klimetschek [mailto:aklimets@adobe.com]=20
Gesendet: Dienstag, 25. Oktober 2011 23:20
An: users@jackrabbit.apache.org
Betreff: Re: AW: How do I get the total size of nodes/rows from a =
QueryResult

On 25.10.11 18:06, "Sascha.Theves@bosch-si.com"
<Sascha.Theves@bosch-si.com> wrote:

>> Use the getSize() method of the returned node or row iterator.
>
>I think that returns the size of elements in the iterator and that is
>hopefully not the complete size if I used the limit and offset =
options...

Yes, I also think it is the actual result "window" size (if given), but
not the total. (But I am not 100% sure).

>I used an XPath query before switching to jackrabbit 2.2.9 but I =
thought
>I should change my query to SQL2 because xpath is deprecated now.

It is deprecated in the spec - but most Jackrabbit based systems use =
xpath
heavily so it is not going away soon. SQL2 on the other hand is still =
very
young - and while having some new features, such as joins, it is not as
optimized and proven as Xpath or SQL1.

(We should really put that in a FAQ).

>So what is your recommendation now? Should I switch to SQL2? What about
>the total size then? Or should I still use xpath? I am currently
>switching because it seems that my xpath queries do not work any =
longer.
>A lot of tests do fail now... I have to track this down but it seems =
that
>I run into another backward incompatibility issue here...

I guess it's only that getSize() returns -1 for you now in most cases =
(if
you come from Jackrabbit 1.x), as that is an optimization in the engine.
To force a result, use the orderby @jcr:score trick mentioned by Ard.

You can't be using limit & offset with Jackrabbit 1.x and Xpath...

Cheers,
Alex

--=20
Alexander Klimetschek
Developer // Adobe (Day) // Berlin - Basel