Hi Marcel,
Thanks for the quick response. I implemented your test - creating the
repository in 1.1.1, inserting the data and querying the data - with an
answer of 30 nodes. Upgraded the repository to the 1.2, queried the data
and indeed did get an answer of 30 nodes. Just for clarification, this was
a JCA rar running in JBoss with a JNDI Datasource using a MySQL Database.
So for this case, everything seems to be working fine. Nevertheless, in my
case of a much larger repository, the upgrade did not seem to work. I have
also tried deleting both the repository and workspace index directories, and
still get all nodes rather than only those matching the date range.
I will do further tests, if there are suggestions to where to look, I will
give them a try. Also, I am willing to look into the code and/or run it
through a debugger, if you can give me some direction where to look.
Thanks,
-Dave
On 1/12/07, Marcel Reutegger <[EMAIL PROTECTED]> wrote:
Hi David,
David B. Johnson wrote:
> I am running a repository on the 1.2 Release Candidate that I upgraded
>>from 1.1.1, i.e., I did not reload the data, I am using the same
> repository. It is a JNDI DataSource Repository with a MySQL backend,
> running as a JCA rar in JBoss. I am using an SQL query that was working
> fine in 1.1.1 and now appears to not work in 1.2, specifically the query
> is of the form:
>
> select * from Story where jcr:path like '/News/Stories/%' and
> publishDate >= DATE '2006-11-01' and publishDate <= DATE '2006-11-15'
>
> Instead of only returning stories with a publishDate within the date
> range, the system appears to return all stories.
>
> Do I need to create a fresh repository, and re-import the data?
no, that shouldn't be necessary. all changes in 1.2 are backward
compatible. at
least that's the goal.
I quickly tried to reproduce the issue with a simple test case [1]. First
I ran
the test on a jackrabbit 1.1 and then again on a jackrabbit 1.2 pointing
to the
existing 1.1 repository data. Both returned the same 30 results.
> If I can help with any additional information, please feel free to
> contact me.
Is there anything else you changed when you upgraded jackrabbit?
Can you please check if you can reproduce the issue with the below test
case in
your environment?
regards
marcel
[1] test class:
package org.apache.jackrabbit.core.query;
import javax.jcr.RepositoryException;
import javax.jcr.Node;
import javax.jcr.query.QueryManager;
import javax.jcr.query.Query;
import java.util.Calendar;
import java.text.SimpleDateFormat;
/**
* <code>DateRangeTest</code>...
*/
public class DateRangeTest extends AbstractQueryTest {
private static final SimpleDateFormat FORMAT = new
SimpleDateFormat("yyyy-MM-dd");
public void testRange() throws RepositoryException {
if (!superuser.itemExists("/data")) {
System.out.println("creating test data...");
Node data = superuser.getRootNode().addNode("data");
for (int i = 0; i < 100; i++) {
Calendar c = Calendar.getInstance();
c.add(Calendar.DATE, i);
data.addNode("node" + i).setProperty("date", c);
}
superuser.save();
}
QueryManager qm = superuser.getWorkspace().getQueryManager();
Calendar lower = Calendar.getInstance();
lower.add(Calendar.DATE, 20);
Calendar upper = Calendar.getInstance();
upper.add(Calendar.DATE, 50);
String sql = "select * from nt:unstructured " +
"where " +
"jcr:path like '/data/%' and " +
"date >= DATE '" + FORMAT.format(lower.getTime()) + "' "
+
"and " +
"date <= DATE '" + FORMAT.format(upper.getTime()) + "'";
Query q = qm.createQuery(sql, Query.SQL);
System.out.println("result size: " + q.execute
().getNodes().getSize());
}
}
