Last dimesion bug(error) on multi dimesion.

[SONIC]

Last dimesion bug(error) on multi dimesion.

hi.
first of all. i'm sorry about my poor english. i'm korean.
i made a Anlysis Program with pivot cube.

this pivot cube had 4 dimesions and 6 measures.
but, THE last dimesion show different items count every time.
sometimes, show me 2 items(this is right). but, sometimes show me 3 items.

look at below.

this is wrong item count...
items are just BK & EA..but..THIS LAST DIMESION MAKE 2 BK...THIS IS GABAGE DATA.


this is right...





i already searched your web site's forum.
i'd changed emptyItems property 'deiRemoveFrom cube'. but, not working.

is it bug? can i fix it?

i'm using Bde in Delphi5, iSeries DB2.
PLZ. I NEED YOUR HELP.

[AlexZencovich]

Hi,

How did you add items to that dimension? By separate dataset or it was created from facts dataset?

Does that problem happens on same cube or that happens on different cube builds?

[SONIC]

I'd made a Sql Statements with "group by & order by" clause.(sql statement has no problem.)
and then Add that statements to TQuery(BDE) Component's Sql Property.

TPivotCube's FactTableDataSet propety is TQuery(BDE).
(TPivotCube is only one. TQuery is Only one.)



I'd Setup Dimesion every item like below.
(chose DataSet(Query1) and then Chose FieldName.)...



and then i'd just put Code in OnButtonClick Event like this.

PivotCube1.Active := false;
PivotCube1.Active := true;

[AlexZencovich]

1 Try to create a separate query for that dimension with sql statement like
select distinct YYYY from Facts

2. Does problem happens randomly on same cube or it just happens on different cube builds?

PS
You do not need to do "group by" or "order by" for build cube, it just take extra SQL server time for performing.

[SONIC]

thanks for your help.

1. i deleted 'order by'. but, i can't delete 'group by'. I need this.
i tried to add 'distinct'. but, it's not working.

2. this problem happens randomly on same cube.
i tried to connect New TPivotCube..but. it is same...

[AlexZencovich]

I could not reproduce that problem in anyway. I hope may be someone with similar problem will able to provide more information. But for now I never heard about that problem with PivotCube.

[SONIC]

i found what problem is.

first of all... thanks for your help.

i made Sql statements first like this.

========================================
select
fld1, fld2,
case when fld1 ='1' then 'EA'
when fld2 = '0' then 'EA'
else 'BK' end ckst

from SomeTable......
========================================

what can i explain this(red).
i used 'Case' Clause. i think that PivotCube might confused User defined field.

and then i changed this field.
deleted 'Case' Clause and used other way.

now. my project is ok. My project shows me right data everytime.

thanks for help.

[AlexZencovich]

PivotCube has no any attention to sql clauses. it does not use it in any way. That may be any data stream, the only it have to be TDataSet or ADO _RecordSet compatible (support EOF, First, Next etc methods)
I just suppose usage to 'EA' may have differently treatened by SQL server and produce 2 EA (which looks similar byt have different locales for example) so in CompareVariant routine EA(fld1) <> EA(fld2)