10g Dimension Hierarchy with strange behaviour

I have a Dimension Table (Calendar) with the following columns:
Date------Year------Semester-----Desc_Sem-------Trimester-------Desc_Trimester--------Month------Desc_Month-------Day
20110101---2011--------1-----------1st Semester----------1----------------1st Trimester---------1-------------January---------1
20111231---2011--------2-----------2nd Semester----------4---------------4th Trimester--------12-------------December--------31
My hierachy looks like this: !http://imageshack.us/photo/my-images/851/bmm.jpg/!
What happens is that when i drill down everything is ok until the Month level but when i try to drill down from the Month level to the Day level it only shows results if i drill down on February!!!
!http://imageshack.us/photo/my-images/830/drilln.jpg/!
A bit more information to add here:
- If on the hierarchy level "Month" instead of "Desc_Month" as the logical level key i choose "Month" (1,2...12) than i have no problems with the drill and it works perfectly...
- I tried the same thing on the samplesales.rpd and the exactly same thing happens. The hierachy there is Year->Half->Quarter->Month->Week->Day and when we drill from Month to Week we also get no results.
Any ideas here? I've searched for some time now and haven't seen anything that can help me here, to be honest i didn't even find a problem similar to this one.
Thanks

Not sure if this is the cause, but your hierarchy isn't strictly correct.
In a hierarchy, the key of each node needs to relate to only one parent. You have Day 1 which could belong to January, February etc. This isn't a key, it's an attribute of the detail level. Likewise, your semester doesn't belong to one year, your quarter doesn't belong to one semester etc.
With the data you have, the 'hierarchy' should be very flat:
[Grand Total]
--[Year]
----[Detail] (key is date, and day is an attribute)
--[Semester]
----[Detail]
--[Trimester]
----[Detail]
--[Month]
----[Detail]
You could define drills paths between levels to sort out your navigation.
If you want your original hierarchy, then your calendar table needs key fields to identify the parentage.
Eg Semester key is 2011-1, 2011-2. Description is 1st Semester.
This is especially important if you move to 11g and want to use the hierarchy features.
Regards,
Robert
Edited by: Robert Tooker on Nov 25, 2011 2:29 AM

Similar Messages

Star Schema Dimension Hierarchy Changes Over Time

Say you have a dimension hierarchy with the level Department that has a child level Division. The Sales Division belonged to the Financials Department up until July 9th, 2011, after which it was moved to the Marketing Department. How can you make it so that both Sales Divisions are interpreted as the same Division, rather than two different Divisions as BI Answers does by default?

I don't think SCD1 would work as BI Answers essentially splits things via both keys and description fields, resulting in odd behavior such as this:
Time
Division
# Employees
All Time
Sales
16
22
Regarding SCD2:
So is there something I need to do on the repository so that BI Answers knows to switch between the two versions of the same Division at that particular time, such that I can make an Analysis such as this:
Month
Division
# Employees
% Change
May
Sales
15
-6%
June
Sales
15
0%
July
Sales
14
-7%
August
Sales
22
+57%
Instead of this:
Month
Division
# Employees
% Change
May
Sales
15
-10%
June
Sales
15
0%
July
Sales
2
14
-87%
0%
August
Sales
22
+57%

Inserting a level in existing dimension hierarchy

hi,
I have to insert a new level in existing dimension hierarchy with out affecting other reports.
Please let me know how to do
Thanks in advance.

I fixed this problem by doing two things. Firstly I deleted
the summary slide at the end of the quiz. This stopped the error
dialog appearing, but the new slides I was trying to insert were
still overwriting an existing question. I then deleted the
questions that were being overwritten, and all subsequent inserted
questions worked correctly. Seems like something got corrupted in
the question slide.
The quiz result summary page gets regenerated so there's no
problem with deleting it if things get corrupted behind the
scenes.

Strange behaviour when using connection pooling with proxy authentication

All
I have developed an ASP.NET 1.1 Web application that uses ODP.NET 9.2.0.4 accessing Oracle Database 8i (which is to be upgraded to 10g in the coming months). I have enabled connection pooling and implemented proxy authentication.
I am observing a strange behaviour in the live environment. If two users (User 1 and User 2) are executing SQL statements at the same time (concurrent threads in IIS), the following is occurring:
* User 1 opens a new connection, executes a SELECT statement, and closes this connection. The audit log, which uses the USER function, shows User 1 executed this statement.
* User 2 opens the same connection (before it is released to the connection pool?), excutes an INSERT statement, and closes this connection. The audit log shows User 1, not User 2, executed this statement.
Is this a known issue when using connection pooling with proxy authentication? I appreciate your help.
Regards,
Chris

Hi Chris,
I tried to reproduce your complaint, but was unable to. I didnt use auditting however, just a series of "select user from dual" with proxy authentication. You might want to see if you can put together a small complete testcase for this and open a sr with support.
Cheers
Greg

Value based dimension hierarchy in OBIEE 10G

Hi,
I am a user of OBIEE10G. I have to create a value based dimension hierarchy in my RPD. My hierarchy should be such that when I click on any VP Name, the respective employees working under him/her should be shown. But my columns of VP Name and Employee_Name are present in separate dimensions. Hence when I drag the column of VP Name to Employee dimension and then create the dimension hierarchy, I get an error of illegal cross join .
Is value based hierarchy possible in 10G?
Pls reply.
Thanks
Ankita
Edited by: 848789 on Sep 28, 2011 8:40 PM

Hi Ankita,
As per my Understanding the VP column did't map to any LTS .
When ever u have multiple LTS in Dim Admin tool won't create Dimension Hierarchy manually we need to create that.
Follow these steps to resolve your Problem...
1) give the simple join in physical layer for that two tables.
2) Create one Dim table in to BMM layer drag all columns in 2 tables in to BMM Dim table so u can able to see all the columns is mapped in to their respective LTS.
3) Now u can create hierarchy manually step by step.Like Grand total Parent ----> Child level .
4) If u have any fact table joined with these dimension in physical layer give the logical joins in BMM layer and set the content levels properly.
Let me know need any detail clarification.
Thanks,
Satya

Strange behaviour with nested overflow headers

Hi there,
I have a strange behaviour with nested overflow headers. I have a hierarchy with some dynamic tables an several headers for the hierarchy steps. When the table entries lead to an page overflow, I would assume, that all parent headers will be shown. But unfortunatley only the last (the deepest in the hierarchy) defined overflow header will be shown. Here you can see a quick example of it. (http://www.nemonon.de/HeaderProblem.xdp)
Any help would be very welcome.
Thanks and greetings,
Joerg

That will sure work for some requirements and thanks for the ideas. But this would fail for the second table. Perhaps my example xdp only shows the behaviour of the nested overflow headers and it´s structure is reduced to show that. The real situation is a bit different. The parent containers also contain some other content. The Problem is not the table...I've just chosen it to produce some content and another nested header. Just try to uncheck the overflow header of the table. This will lead to, that the next defined header (in this case "HeaderThree" is shown. When you go on by also disabling that header, the next one will be shown...and so on. This shows, that the headers for their own are functional, but they fail in combination.
Greetings,
Joerg

View Master mode strange behaviour with jpg files

Hy all,
I'm using Aperture, and I'm tweakin a lot the SW to discover all secrets ....
I'm really concerned about a strange behaviour of the SW in view mode.
If I open a jpg file, do not apply ANY modification, and apply the "show master" command (pressing M), there's a subtle behaviour :
*the images get sharper !!!!*
It behaves as if an "edge sharpening" was applied !!!!!!
With _RAW files this thing doesn't happen at all._
Even stranger, if I zoom to full definition (by pressing Z), and toggle between normal view and "show master" view, the immage remains the same !
The effect is not so evident with all images, but sometimes it is !!
Anyway, I expect not to have such kind of things with a professional SW.
With professional cameras and good quality files it's really annoying not to beeing sure what's the REAL content of the jpg image.
Has anyone noted this thing ? Is there a reasonable explanation ?
The only one I tried to figure out is the following : may be Aperture applies some kind of resampling algorithm with jpg files (bicubic, bilinear, etc).
May be this resampling is not applied with 100% magnification, as weel as with raw files.
But it's only my guessing. Now, I'm tweaking other viewers and other SW to understand what's the real jpg image stored in the file : the blurer one form normal view or the sharper one with "show master" mode.
I stress once again : of course, the jpg file I'm talking about have no settings at all, just imported and toggled with M button.
Thanks in advance.
Regards,
Enzo

Hy,
I'm afraid I'm not been clear.
The problem is this :
1) jpg image opened in normal view mode, no zoom, (no quick preview, of course) : image with a certain amount of blour
2) jpg image opened in show master mode, no zoom, (no quick previre, of course) : image SHARPER with respect with the previous.
No edit applied from import, so "show master" command shouln't apply any change in rendering, should it ?
The only thing I said on 100% view is that such difference (between normal mode e show mastermode) is not present with 100% magnification.
That was the issue.
Enzo

Strange behaviour with Safari.

Very strange behaviour with Safari this morning. Pages loading piecemeal or partially. The content only becomes viewable after the cursor has passed over the area. Have cleared caches and reported it but I'm wondering if this is a local matter here in the UK. Anyone suffering similar problems? Firefox is responding perfectly normally.

I see a reply has come back via a previous post. As I am trying to get a response to this one I shan't reply there but I should point out that for me, this is an isolated problem. I do use open DNS but with no track history of problems there , I can't blame that.
In the last few minutes normality has returned.

IPhoto: strange behaviour with preview thumbnails

Is there a simple way in iPhoto 8 to rebuild the thumbnail database?
I recently noticed a strange behaviour iPhoto 8 has with previews: In some of my event categories, iPhoto does not display the correct preview thumbnail of the picture, but a preview thumbnail of another event. When I click on the thumbnail, it reloads the appropriate preview and it does not happen again - with that picture. Also, if I double click an event to open it, the small thumbnail of the event's key photo that is to be displayed in the title bar (next to the All Events button) is not always correct.
Thank you for your help!
Stefan

I have similar problems (lately) with iPhoto 6.0.6 -- thumbnails do not equal the images they 'thumbnail' of.
Sometimes when it opens, it (recently) asked me to rebuild the thumbnails, but not always.
Can I force a thumbnail rebuild using the command you gave - Apple-option while opening iPhoto? In iPhoto 6.0.6 this primarily forces a new selection of a library.

How to delete duplicates in oracle 10g ( strange behaviour)

Recently we migrated from oracel 8i to oralce 10g and we face this problem.
When we try to delete duplicates using rowid ( analytical functions row_number()/
or normal delete ) and commit the same , still we find some duplicates to be existing because of which we are not able to enable costriants and resulting in process failure.
When we run the same delete statement next time it removes more duplicates than the required or sometimes required duplicates resulting in abnormal behaviour.
I don not understand this strange behaviour after upgrading to oracle 10g.
It'd be great if some one who has idea on this can throw light on the same.
thanks

Gasparotto,
Thanks a lot for letting me know a new procedure to delete duplicates sing lead function.
I've tried this code on a temp table and it worked, let me use the same query on the prodn side and test the same.
Procedure for deletion of duplicates using lead analytical function.
create table temp ( col1 number(2) , col2 number(2) , col3 number(2));
insert into temp values ( 1,2,10);
insert into temp values ( 1,2,20);
insert into temp values ( 1,2,30);
insert into temp values ( 3,2,10);
insert into temp values ( 3,4,12);
insert into temp values ( 3,4,45);
commit;
COL1 COL2 COL3
1 2 10
1 2 20
1 2 30
3 2 10
3 4 12
3 4 45
select col1,col2 , col3, LEAD(rowid) OVER (PARTITION BY col1,col2 order by null) from temp;
COL1 COL2 COL3 LEAD(ROWID)OVER(PA
1 2 10 AAAVBjAApAAAFyGAAB
1 2 20 AAAVBjAApAAAFyGAAC
1 2 30
3 2 10
3 4 12 AAAVBjAApAAAFyGAAF
3 4 45
6 rows selected.
select rowid , temp.* from temp ;
ROWID COL1 COL2 COL3
AAAVBjAApAAAFyGAAA 1 2 10
AAAVBjAApAAAFyGAAB 1 2 20
AAAVBjAApAAAFyGAAC 1 2 30
AAAVBjAApAAAFyGAAD 3 2 10
AAAVBjAApAAAFyGAAE 3 4 12
AAAVBjAApAAAFyGAAF 3 4 45
SQL> DELETE temp
WHERE rowid IN
( SELECT LEAD(rowid) OVER (PARTITION BY col1, col2 ORDER BY null)
FROM temp ); 2 3 4
3 rows deleted.
SQL> select rowid , temp.* from temp ;
ROWID COL1 COL2 COL3
AAAVBjAApAAAFyGAAA 1 2 10
AAAVBjAApAAAFyGAAD 3 2 10
AAAVBjAApAAAFyGAAE 3 4 12
Thanks for the reply

Total levels on pivot with dimension hierarchy column incorrect

Hi,
I have a pivot table based on a single subject area, based on an Essbase cube.
The problem I have is that some of the totals against the collapsed dimension hierarchy column are incorrect, it is as if they remember the pre-updated state, even though I have bounced the services and turned all caching off.
When I expand the node then the totals appear correctly, but when reverted to the collapsed state it shows nothing again.
I have turned off caching via EM.
I have tried all flavours of totals, default, server complex aggregate et al with the same result for all.
I have double checked the results at source with an identical query in SmartView and the upper level dimension total is populated in its own right, as well as its children carrying totals which do not sum back to zero.
Any suggestions please??
thanks,
Robert.

Can you validate the query hitting to the database and compare the results?
Mark if helps.
Thanks,

Strange behaviour with Slideshow placement since upgrade to 3.2 from 3.1

Hello, since I upgrade from 3.1 to 3.2, I have some strange behaviours but in particular with the placement of a slideshow on a new page. It is not displaying in the same position as I have in in Design. It is way off and going up over my menus when I look at it in Safari or Chrome. Does anyone know how to fix this or could it be a bug?

Ben,
When you upgrade, only apex_030200 will be created and that will use the tablespace you specify. It does not have to be the same one used by flows_030100.
Scott

Strange behaviour with smart music list in iTunes

I don't know how is the name of this list in english. In portuguese iTunes calls "Lista Inteligente". So, I eddited a list to show all musics never played up to 100 months ago. If I click on a music in this list, before it plays, the music disappear from the list as it would be played and nothing will be played. And so one, the next and etc. No date is shown in "last play" but the music isn't anymore to see in this List.
This list works well for many years but suddenly I have this strange behaviour.
could it be something wrong with new update? Now I have iTunes installed 11.3.0.54

Click the gray disclosure triangle in front of the iPod touch name.
Thus I cannot do any functions with the music on my iPod - such as delete songs, update song info, etc.
I would suggest making any changes to songs in iTunes and then transferring the songs to the touch.
You can create playlists and autosync on lythos playlists.
Then add/delete from those playlists, update items in iTunes then plug in the touch to sync it.

KYF exhibits strange behaviour with decimals

Hi everybody,
a DEC-key figure (IOBJ maintenance: additional properties->BEx->decimal places: Nothing Defined) exhibits strange behaviour regarding the treatment of decimals.
When looking at the data in the cube (InfoCube maintenance->content) the value is 2,086.
In the query the result is 2.0900000000 (%).
Now comes the problem: in BPS (in the buffer) the value is 2,08. The system simply truncates the figure! This is the value which is also used in the FOX and produces the wrong result.
I need the system to sensibly round the values correctly to 2 decimals OR work with an additional decimal place. How can this be done?
Thanks for any input,
Martin

Hi,
In the manual planning, while designing the layout, there is a setting in data columns tab called decimals. You have to insert 2 under this. Then in the layout data will be displayed with 2 decimal places.
Hope this helps.
Regards,
Isha

Strange behaviour msi 7236 with e7400

Hello,
Is the msi-7236 supporting the e7XXX Intel c2d family ?
I've just switched from an e6400 to an e7400 and I've noticed some strange behaviours, mainly
the fact that the performances are the same with both processors ( and even a bit reduced with
the e7400 ! ).
Another big problem is in the bios CPU configuration section: instead of 6 options + C1E hidden features
there is only 2 + the C1E hidden features. Consequence is simple: impossible to setup Intel speedstep or
the other common options.
Additionally there is also a problem when displaying the frequency at the boot-time ( it displays 346xx MHz !)
and in cpuz, the coefficient displayed is 6 ( instead of 8 I guess for this entry level cd2 model)
As the latest msi 7236 bios is now 2 years old (v3.7), I guess this is the motherboard which is
responsible for this weird behaviour.
So anyone can help me with theses questions ?

945P Neo3, bios is version 3.7, to be accurate, internal nomenclature is 'MS 7236 revision 2.2'.
With ChristalCPUID I've succeed into putting the coefficient @10.5 so now the performances of the e7400 are as I was expecting, but the bios is still initializing the CPU with the lowest coeff.
Advanced features (64 bits instructions, SSE4 ,speedstep etc) are still not recognized.
I've also noticed that the e7400 is not listed on the 'officially supported CPU list' motherboard webpage.
How can it work partially without system failure ?
Would it be possible that 45nm are supported by this (old) MotherBoard even if it's not clearly specified by MSI ?

10g Dimension Hierarchy with strange behaviour

Similar Messages

Maybe you are looking for