Mailing-List: contact abdera-user-help@incubator.apache.org; run by ezmlm
Precedence: bulk
Reply-To: abdera-user@incubator.apache.org
Received-SPF: pass (athena.apache.org: local policy)
Reply-To: <remy@arrova.ca>
From: "Remy Gendron" <remy@arrova.ca>
To: <abdera-user@incubator.apache.org>
Cc: =?iso-8859-1?Q?'R=E9my_Beaumont'?= <remy.beaumont@lacapitale.com>,
	"'Martin Guay'" <martin.guay@lacapitale.com>,
	<jcaron@elic.ca>
Subject: Need hook in AbstractEntityCollectionAdapter for paging
Date: Tue, 20 May 2008 14:00:45 -0400
Organization: Le Groupe Arrova Inc.
Message-ID: <681E8623E3FF409C9FD53EE2A5BA0ACD@Arrova001>
MIME-Version: 1.0
Content-Type: multipart/alternative;
	boundary="----=_NextPart_000_001D_01C8BA81.E7FFD5F0"
Thread-Index: Aci6o21I1S4GfJCgRC28pQVY9DuolQ==

------=_NextPart_000_001D_01C8BA81.E7FFD5F0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable

Hello all,

=20

To make paging as much an orthogonal feature as possible, w/r to the
application code, I am not using the very good proposition from our =
friends
Chris and Bryon
(http://www.mail-archive.com/abdera-user@incubator.apache.org/msg00525.ht=
ml)
. I may not always have to possibility to use an update sequence as I =
will
quite often work on legacy/existing database schemas.

=20

I am relying instead on the native paging features of the underlying
database, not requiring any injection into the SQL where clause. For
instance, with Oracle I have this call to add paging:

=20

    /**

     * @see
http://www.oracle.com/technology/oramag/oracle/06-sep/o56asktom.html

     */

    @Override

    protected void addPaging(StringBuilder sql, MapSqlParameterSource
params, ServiceContext context) {

        params.addValue("startIndex", context.getStartIndex());=20

        params.addValue("endIndex", context.getStartIndex() +
context.getMaxResults() - 1);

=20

        String prefix =3D "select * from ( select /*+ FIRST_ROWS(" +
context.getMaxResults() + ") */ a.*, rownum r from ( ";

       =20

        sql.insert(0, prefix);

        sql.append(" ) a where rownum <=3D :endIndex ) where r >=3D
:startIndex");

    }

=20

My request:

=20

I follow Tom Kyte (asktom) recommendation about paging. My Abdera glue =
code
alters my service layer context so that the application code believes it
must retrieve n+1 entries instead of n. That way, I can easily know =
(with a
very low database cost) if there is the need for a next page link to be
inserted into the feed.

=20

I have overridden AbstractEntityCollectionAdapter.addFeedDetails() to =
insert
the paging links if necessary.

=20

However, the FOM for a feed does not allow to remove entries from the
collection (Feed.getEntries()). Remember, my glue code had the adapter
retrieve one too many entries in getEntries().

=20

Hence, I would need an afterGetEntries(request, entries) called in the
addFeedDetails() prior to the iteration to alter the returned pojo =
entries
(in my case remove the additional entry).

=20

I don=92t want to require the application developer to call some method =
in his
getEntries().

=20

For now, I have copied and modified addFeedDetails() in its entirety in =
my
AbstractEntittyCollectionAdapter class to implement the afterGetEntries
hook.

=20

Can anyone recommend a better approach? If not, I will enter a JIRA for =
this
request.

=20

Thank you again,

=20

R=E9my


------=_NextPart_000_001D_01C8BA81.E7FFD5F0--