[Mondrian] Problem with aggregate tables..

Julian Hyde jhyde at pentaho.org
Sun Feb 17 14:19:04 EST 2008

I don't think any progress has been made on this issue. I looked at the
history of the email thread, and it quickly diverged into another issue. I
ran the tests, but as I said, I need that regression test, and I'm not going
to have any time soon to write it. Maybe someone else will write it - my
description re. TestAggregationManager is as much as I'll give as far as
'pointers' - any more and I'd be writing the code myself - but I have a
feeling that no one else will write the test.
So, please log a bug with all of this information in it, so it doesn't get
forgotten, and we will get to it one day.


From: mondrian-bounces at pentaho.org [mailto:mondrian-bounces at pentaho.org] On
Behalf Of Ati Rosselet
Sent: Sunday, February 17, 2008 5:04 AM
To: Mondrian developer mailing list
Subject: Re: [Mondrian] Problem with aggregate tables..

Julian (and others?)
Ok... I give up.. I have no idea how to properly test these changes.  Are
they still applicable to the current CVS version? (I assume so.. since I
haven't seen any notes/commits pertaining to this problem).   You mentioned
that you were going to run it through the regression suite to see if any
problems popped up - but I never heard of any results.. Does this mean that
basically all is ok?  This fix is kind of important.. pretty vital to us in
fact, and I'm surprised no-one else has encountered the problem I address in
it.   If you (or anyone) would be willing to either write the test, or give
me some pointers (I'm completely lost - probably partly from being in the
middle of a large .NET project :)), I'd be very grateful.   I'll try to get
the CVS version of 3.0 and see if the patches can still be applied.


2007/10/23 Julian Hyde <julianhyde at speakeasy.net>:

Thanks for the fix. I've run your changes through the regression suite.
Looking good so far. I'll run the suite in all of its various property
settings tonight and see if anything pops out.
Since this change is a bug fix, I need a regression test before I will check
in. I think you could construct a testcase based on the foodmart database.
Add the test to TestAggregationManager; testMultipleMeasures is an example
of a test which runs an MDX query and checks that a particular SQL query is
generated. You will also need to use TestContext.create() to alter the XML;
testKeyExpressionCardinalityCache is an example of a test which does that.


From: mondrian-bounces at pentaho.org [mailto:mondrian-bounces at pentaho.org] On
Behalf Of Ati Rosselet

Sent: Monday, October 22, 2007 5:23 PM
To: Mondrian developer mailing list
Subject: Re: [Mondrian] Problem with aggregate tables..

sorry... cut and paste got me.. end of first function ended up after the end
of the 2nd modified function. so

         * Finds ALL the child tables of the fact table with the given
         * used in their left join condition. This is used by the
         * while characterizing the fact table columns. (previous only
returned FIRST child)
       public ArrayList<RolapStar.Table>
findTablesWithLeftJoinCondition(final String columnName) 

            java.util.ArrayList<RolapStar.Table> tables = new
            for (Table child : getChildren()) {
                Condition condition = child.joinCondition;
                if (condition != null) {
                    if (condition.left instanceof MondrianDef.Column) {
                        MondrianDef.Column mcolumn =
                            ( MondrianDef.Column) condition.left;
                        if (mcolumn.name.equals(columnName)) {
            return (tables.size()==0?null:tables);

and  remove the return part from other function..


Mondrian mailing list
Mondrian at pentaho.org

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.pentaho.org/pipermail/mondrian/attachments/20080217/43be873f/attachment.html 

More information about the Mondrian mailing list