ORA-36836 about cube dimensionality ???

Hi there,
I've been trying lately to use AWM to build an OLAP cube based on some sample data I have related to phone communications. Up to now I've been able although painfuly to build four dimensions (date/time, a-number, b-number, exit cause) and a cube with measures "number of calls" and "total call duration". Currently everything seems to load correctly (dimensions and cube), however when I try to view the cube data I get the following message :
ORA-36836: The dimensionality of object OLAP.TRAFFIC is different than the view token
This doesn't make any sense to me, and I've not been able to find any explanation on the web, except for those two lines :
Cause: The dimensionality of the cube does not match the view token
Action: Check view token for unrelated hierarchies clauses
Well the remedy looks even worse than the disease, since this is even less understandable than the former.
Does anyone has a clue about what might be wrong here, or what to do to go forward ?
I'm using the free Oracle 11g on CentOS, I guess this must be the XE edition, and AWM was part of Oracle Client on Windows XP. This piece of software seems to be riddle with bugs, data not updating properly after changes, objects only partly deleted after a delete command, lots of "uncommitable" and java exceptions of all kinds... I've had several times to totally delete the whole aw and restart from scratch because AWM would not finish deleting some dimension of attribute. A real pain in the neck. Is this related to the fact that I'm using a free version and don't have access to patches ? Well I really wonder how can anyone do some serious development with this tool. Or am I missing something ?
Thanks for your help,
Christian

Hi again Stuart and all,
I'm back after some patching... :-) I've applied the following patches to my installation :
On the server side patches 7375639 (Oracle Critical Patch Update October 2008) and 6992921 (OLAP B PATCH FOR 11.1.0.6).
On the client side I've upgraded to AWM 11.1.0.7.0A
And ... tadadam ... I still have the same error message after recreating the whole workspace from scratch :-(
Here is a transcript of the error message I get :
[edit many similar lines]
oracle.dss.dataSource.common.OLAPException: BIB-9009 Oracle OLAP could not create cursor.
oracle.express.idl.util.OlapiException: java.sql.SQLException: ORA-36836: The dimensionality of object OLAP.TRAFFIC is different than the view token
ORA-06512: at "SYS.GENCURSORMANAGERINTERFACE", line 42
ORA-06512: at line 1
oracle.express.idl.util.OlapiException: java.sql.SQLException: ORA-36836: The dimensionality of object OLAP.TRAFFIC is different than the view token
ORA-06512: at "SYS.GENCURSORMANAGERINTERFACE", line 42
ORA-06512: at line 1
at oracle.dss.dataSource.common.QueryDataDirector.addDataDirectorListener(QueryDataDirector.java:795)
at oracle.dss.dataView.ModelAdapter.setDataDirector(ModelAdapter.java:173)
at oracle.dss.crosstab.CrosstabModelAdapter.setDataSource(CrosstabModelAdapter.java:54)
at oracle.dss.dataView.Dataview.setDataSource(Dataview.java:396)
at oracle.olap.awm.dataobject.dialog.olapi.UBiBeanUtil.getCrosstab(Unknown Source)
at oracle.olap.awm.dataobject.dialog.olapi.UBiBeanUtil.showCrosstab(Unknown Source)
at oracle.olap.awm.dataobject.dialog.olapi.UBiBeanUtil.showData(Unknown Source)
[edit many similar lines]
I'm able to display the dimension views and the cube view and they "look" ok, but I get this error when I right-click to "View data <MY CUBE>..." or "View data <ONE OF THE MEASURES>..."
So it looks like there's something in the way my data is wrapped up that really bothers the Oracle OLAP engine. What I'm trying to do is to migrate a cube that works on Mondrian to Oracle OLAP, so it means the problem is probably related to some specific requirement that the Oracle OLAP engine imposes on the way the data is presented.
In any case the error messages are totally meaningless to me. Does anyone has an idea what the "view token" is ?
Thanks a lot for your help,
Christian

Similar Messages

About cube

currently, i am doing a project using xmii, and i use olap query to visit data in bw, but i am a little confused about that query .
so could someone tell me what is a cube ,what is the structure of a cube

Hi,
what is a cube ,what is the structure of a cube
Cube is a multi dimensional object which has designed using extended star schema,
cube is surrounded by 16 dimension tables, these dimension tables are inturn connected with SID tables which are linked with master data tables,
dimension IDs are generated when transaction data is loaded into CUBE, these dimension ids are generated based on SID.
SIDs are generated when master data is loaded.
Thanks,
Pavan.

Getting ORA-03001 on cube query

I am getting a ORA-03001 error on the following query - I am utlizing a cube query and the main data view (subquery) is using a connect by. I believe it is the connect by which is causing the problem as I have tried the query with the table data (BASEDATA) and it works fine. I also tried removing the having clause, and that worked. Looks like a bug. I am on 10gR2.
SELECT progm.attrib_id,NH.workstream1_parent_id, nh.cost_type, nh.cost_benefit, eh.expense_majorgroup_id, eh.expense_group_id, b.year_,b.period_id, b.measure, data_type_id,
Sum(Value) value
,grouping_id( b.cost_centre_id, NH.workstream1_parent_id, progm.attrib_id) NH_ID
,grouping_id( b.Expense_type_id, EH.expense_group_id, EH.expense_majorgroup_id) EH_ID
,grouping_id( b.year_,period_id) yRPER_ID
,grouping_id( nh.cost_Type, nh.cost_benefit) Cost_ID
FROM projectmeasure_data B, nomlocationhierarchy NH, nomexpenseshierarchy EH
, (SELECT workstream1_parent_id, wspa.attrib_id FROM workstream1_parent_attrib WSPA inner join
attributes AA ON WSPA.attrib_id=AA.attrib_id WHERE attribType_id=28) Progm
WHERE B.cost_centre_id=nh.cost_centre_id AND
b.expense_type_id=eh.expensetype_id AND
b.measure_id=cm.measure_id AND
progm.workstream1_parent_id=nh.workstream1_parent_id
group BY b.measure, b.data_type_id,
cube ( b.cost_centre_id, nh.workstream1_parent_id, progm.attrib_id,
b.expense_type_id, eh.expense_group_id, eh.expense_majorgroup_id,
b.YEAR_,period_id, nh.cost_type, nh.cost_benefit)
HAVING
(grouping_id( b.cost_centre_id, NH.workstream1_parent_id, progm.attrib_id) IN (4,6,7))
AND
(grouping_id( b.Expense_type_id, EH.expense_group_id, EH.expense_majorgroup_id) IN (4,6,7) )
AND
( grouping_id( b.year_, b.period_id) IN (3,1,0))
AND
(grouping_id( nh.cost_Type, nh.cost_benefit) in (1,0))
where
ProjectMeasure_data is
SELECT
A.Cost_Centre_ID,
A.Expense_Type_ID,
A.Period_ID,
A.Year_,
A.Data_Type_ID,
M.RptSlot AS Measure,
A.Value,
H.Measure_ID
FROM ( SELECT Cost_Centre_ID, Expense_Type_ID, Period_ID, Year_, Data_Type_ID, measure_id, BASEDATA.Value
FROM BASEDATA
WHERE ((Year_=2004 AND Measure_ID=1) or ( Year_ >= 2005)) and BASEDATA.Data_Type_ID in (1,2) ) A , qrycurrentmeasures M,
(select connect_by_root(measure_id) root_id, measure_id, CASE LEVEL WHEN 1 THEN switch_year*100+switch_per ELSE Decode(PRIOR(end_period),13,PRIOR(end_year)+1,PRIOR(end_year))*100+Decode(PRIOR(end_period),13,1,PRIOR(end_period)+1) END switchyrper, PRIOR(end_year)*100+PRIOR(end_period) prior_endyrper, end_year*100+end_period endyrper, LEVEL D
FROM measuretypes
CONNECT BY NOCYCLE PRIOR future_measure_id = measure_id
union
SELECT root_id, measure_id, switchyrper, prior_endyrper, endyrper, D FROM
(SELECT connect_by_root(measure_id) root_id, measure_id, switch_year*100+switch_per switchyrper, Decode(end_period,1,end_year-1,end_year)*100+Decode(end_period,1,13,end_period-1) prior_endyrper, Decode(PRIOR(switch_per),1,PRIOR(switch_year)-1,PRIOR(switch_year))*100+Decode(PRIOR(switch_per),1,13,PRIOR(switch_per)-1) endyrper, LEVEL D
FROM measuretypes
CONNECT BY NOCYCLE PRIOR hist_measure_id = measure_id) WHERE D <> 1
) H
WHERE
(m.measure_id=H.root_id) AND
(H.measure_id=a.measure_id ) AND
( (yearperiod(a.year_,a.period_id) BETWEEN H.switchyrper AND h.endyrper) OR
(h.switchyrper IS NULL AND yearperiod(a.year_,a.period_id) <= h.endyrper) )
I ended up using a temporary table to store the results of projectmeasure_data. Does anyone know a way around this? to me it looks like a bug.

I'm still wondering about this one.

XOQ-01600: OLAP DML error "ORA-33858: 11g Cube

Hi All ,
I was trying to apply the cube changes after changing the sparsity definition against one of the dimension under Storage tab and it is throwing below error .
Any reason ?
An error has occurred on the server
Error class: Express Failure
Server error descriptions:
DPR: cannot create server cursor, Generic at TxsOqDefinitionManager::generic<CommitRoot>
INI: XOQ-01600: OLAP DML error "ORA-33858: The value of the ampersand-substitution expression is NA." while executing DML "SYS.AWXML!R11_MANAGE_CUBE('MARKET_SALES_CUBE_WORKING.CUBE' 'ALTER' 'NUMBER' SYS.AWXML!___R11_LONG_ARG_VALUE(SYS.AWXML!___R11_LONG_ARG_DIM 1) SYS.AWXML!___R11_LONG_ARG_VALUE(SYS.AWXML!___R11_LONG_ARG_DIM 2) 'TIME.DIMENSION' 'TIME.CALENDER.HIERARCHY' 'TIME.MONTHLY.LEVEL' 'COMPRESSED' 'YES' 'YES' 'MARKET_SALES_CUBE_WORKING.SOLVE.AGGREGATIONDEFINITION' 'NO')", Generic at TxsOqStdFormCommand::execute
Thanks in advance,
DxP

I have seen this kind of error show up if you rename an object (e.g. a measure) and then make further modifications. Did that happen in your case? If you export the cube to XML, delete it from AWM, and then recreate it from XML, then it may resolve the problem. If not, and if it is preventing your from making progress, then I would open a service request since we usually need to enable tracing to resolve this class of error.

ORA-01578 about standby database in read only mode

Hi,
I have an ORA-05178 (data block corrupted) about a standby database which in read-only mode. About production database, there is no problem.
After analyze, this is an index segment :
SELECT segment_name , segment_type , owner , tablespace_name
FROM sys.dba_extents
WHERE file_id = 58
AND 218173 BETWEEN block_id and block_id + blocks -1
SEGMENT_NAME SEGMENT_TYPE OWNER TABLESPACE_NAME
SICINHIS01 INDEX MOWIN IDX_DATA01
There is no constraint.
How can I repair this problem ?
Nicolas

Hi,
I have an ORA-05178 (data block corrupted) about a standby database which in read-only mode. About production database, there is no problem.
After analyze, this is an index segment :
SELECT segment_name , segment_type , owner , tablespace_name
FROM sys.dba_extents
WHERE file_id = 58
AND 218173 BETWEEN block_id and block_id + blocks -1
SEGMENT_NAME SEGMENT_TYPE OWNER TABLESPACE_NAME
SICINHIS01 INDEX MOWIN IDX_DATA01
There is no constraint.
How can I repair this problem ?
Nicolas

About Cube design

Hi,
i'd like know if in AWM I can have a single cube with multiple fact tables Fact tables share any dimension, otherwise others are specific of a single fact table (or each fact table needs a different cube) and if a member of the cube can be mapped only by one fact table and not necessarly by both.
Thanks
Giancarlo

I would be able to answer well if you can you describe your facts and elaborate what you want to achieve? In earlier post I was trying to say that suppose you have a fact with dim1,dim2,dim3 with measure1 and you create one cube for it say cube1.
Now do you mean that you have another fact with dimensionality as dim4,dim5,dim6 with measure2? If yes then you should create a second cube with dimensionality as dim4,dim5,dim6 with measure2.
On the other hand if you have second fact as dim1,dim2,dim3 with measure2 then in this case you can have only one cube with two measures and the source to load the cube could look like this
create view cube_source as select dim1,dim2,dim3,sum(measure1),sum(measure2) from(
select dim1,dim2,dim3,measure1,null from fact1
union all
select dim1,dim2,dim3,null,measure2 from fact)
group by dim1,dim2,dim3
Above is my perception but I am not sure if you I really understood your question correctly. Can you elaborate what you are trying to achieve?
Thanks,
Brijesh

Summarized Overview about Cube-Layouts?

Hi experts,
does NetWeaver BI 7.0 provide any kind of summerized overview, which gives information about the number of dimensions existing in all (non-virtual) InfoProviders, their record-count, space-usage, fact-table sizes, etc?!
Or is the only way to get this information the usage of SE16 on a per-table basis?

Hi ,
you can use the functional module.
there are few in se37 . put RSAR* and f4 and cube you will find functional modules which can help you for the issue.
You can find in se38 also few programs with rsar*
You can use sap_infocube_designs functional module also.
Regards
Rahul
Edited by: Rahul Bindroo on Mar 27, 2008 11:57 AM

Request about cube effect

E mann asked me to explain the cube effect I did:
I used PE3.
The cube:
Bring 3 clips into the timeline, each on a separate track.In properties reduce the scale of each one and move them so they are all visible. I found a picture of a cube that I like on the net, copied and brought it into the project as a guide, on a track beneath all the video clips. (Keeping the cube large makes it easier to align the clips)
Take one of the video clips, rotate and drag it to align with a corner of one section of the cube. Apply the corner pin filter to the clip. In the properties window highlight Corner Pin and the handles will appear around the clip. Now just drag each handle until the clip fits the section of the cube. Do this with all 3 clips. If you want a title on the cube, add it.
When you are done, delete the cube photo. Then export the movie.
http://img506.imageshack.us/img506/3401/untitled1cy1.jpg
For the tumbling:
Bring the cube movie you just saved into a new project. Now you can scale down the size of the video cube to whatever suits you. Move the cube to a corner to start. Toggle motion animation on, that's your 1st key frame. Then move the the video ahead X number of frames and drag the frame to a new location and set a new keyframe. Keep doing this till you have the paths complete.
http://img372.imageshack.us/img372/4418/untitled2fd3.jpg
Once you have the path right, set the rotation as you want it for each key frame. Start at the 2nd key frame and type in the amount of rotation you want between key frame 1 & 2. (Note: minus values rotate counter clockwise) If you want it to rotate more than a full turn put in some like -600, when you hit enter it will show as -1X-240.0 which equals 1 full turn counter clockwise plus 240 degrees. (-2X would be 2 full rotations and so on) Fine tune the rotation manually to get it where you want it. Do this at each key frame.
When you're done, export this too. All done. Hope I was clear.

That's why I suggested this thread be started. I would like to hear what others are doing to get new ideas, likely there are many things people are doing which I would never have thought of on my own.
Hopefully people will be willing to share their ideas, and techniques for free.

Function module / table to get information about CUBE & ODS

Hello All,
I would like to compare ODS & Info cube structures between DEV & Quality systems. Could you please let me know is there any best practice available for this activity.
I hope we could do this activity by using FM / reading Data base tables directly, all helpful answers will be rewarded with points.
Regards,
Venkat

Hi Venkat,
The function module BAPI_CUBE_GETDETAIL provides you the detials of infocube (including dimension, characteristics, keyfigures etc).
Moreoverthis FM is RFC enabled; so you can call this FM from DEV and pass RFC connection of PROD to get detials of the cube in prod. This compare the structures between DEV and PROD.
Similarly you can use BAPI_ODSO_GETDETAIL from DSO structure comparision across the system landscape.
Thanks
Vasu

Two dimensional table

Hi all,
I need to know about two dimensional structure in pl/sql?
And how many level of sub queries are optimum for a query?

Note the name of this forum is SQL Developer *(Not for general SQL/PLSQL questions)* (so for issues with the SQL Developer tool). Please post these questions under the dedicated SQL And PL/SQL forum.
Regards,
K.

Read existing cubes with Java OLAP API

Hi,
i am new to the Oracle OLAP API and i have a question for which a do not find the answer in the documentation..
I need to access all existing cubes in a database, get the names, the corresponding dimensions etc.
There is no problem in reading the dimensions their hierarchies, the measures an so on. But i found no method or class to read out the information about cubes (names, dimension).
The only thing i found in the documentation to get a cube was over a MtmMeasureMap, but when i call this method i get the exception: Not supported by server version ORAOLAP10.2.0.1.0
Any hints?
Thanks

In current releases, the OLAP API is primarily a query API. Queries can include measures from multiple cubes, so there is no real purpose for a cube object. Therefore there is no Cube class in the Mdm logical metadata package.
Cubes are useful in OLAP mappings, and there is an MtmCube class in the Mtm mapping metadata package. It is possible to navigate between Mdm and Mtm objects. The code that follows prints out the names of all of the cubes and the names of their dimensions.
I don't understand the "Not supported by server version" error that you got. This sample code works for me in 10.1.0.4, and I expect it would work in 10.2, too.
Geof
MdmMetadataProvider mdp; // given
Hashtable cubes = new Hashtable();
MdmSchema rootSchema = mdp.getRootSchema();
Iterator subSchemas = rootSchema.getSubSchemas().iterator();
collectCubes(rootSchema, cubes);
while (subSchemas.hasNext())
collectCubes((MdmSchema) subSchemas.next(), cubes);
// enumerate cubes
Enumeration cubeList = cubes.keys();
while (cubeList.hasMoreElements())
MtmCube cube = (MtmCube) cubes.get(cubeList.nextElement());
System.out.print(cube.getName() + ": ");
Iterator dimensionMaps = cube.getDimensionMaps().iterator();
boolean first = true;
while (dimensionMaps.hasNext())
MtmDimensionMap dimensionMap = (MtmDimensionMap) dimensionMaps.next();
MdmDimension dimension = (MdmDimension) dimensionMap.getMdmSource();
if (first)
first = false;
else
System.out.print(", ");
System.out.print(dimension.getName());
System.out.println();
void collectCubes(MdmSchema schema, Hashtable cubes)
Iterator measures = schema.getMeasures().iterator();
while (measures.hasNext())
MdmMeasure measure = (MdmMeasure) measures.next();
MtmMeasureMap measureMap = measure.getMeasureMap();
MtmCube cube = measureMap.getCube();
// See if we have already visited this cube for another measure
String id = cube.getID();
if (null == cubes.get(id))
cubes.put(id, cube);
}

How to map a cube to a non-leaf dimension level?

Hi,
I'm using AWM 11.2.0.1.0 and Oracle 11.2.0.2.0 on a CentOS 5.4 64-bit system.
I have a time dimension with levels Year, Month, Day, Hour, Half-Hour and several cubes mapping to the Half-Hour level. Now I'm trying to populate a cube using a time hierarchy Year/Month/Day, i.e., the mapping should be done at the Day level, not Half-Hour.
When adding the time dimension, although I specify that aggregation should be done only on the Year/Month/Day hierarchy, I still see all the levels in the mapping view. And when I try to map the Day level to the fact table, I get the following error:
The transaction is not committable: "Invalid Metadata Objects:
Invalid Object "OLAPIN.SSD_BALANCE.MAP1.DATETIME": "XOQ-01930: The cube dimensionality map has no mapped dimension."What is strange is that this used to work on version 11.1.0.7.0 (I have an example of it right in front of me). Am I supposed to do things differently with the new 11.2 version?
Thanks for your help!
Chris

My guess is that this is a result of
BUG 8264467 - REGRESSION FROM 10G : CANNOT LOAD CUBE DATA AT MULTIPLE LEVELS
This was fixed in 11.2.0.1 on the server and may appear to have nothing to do with your problem. Unfortunately there was a bug in the corresponding change in 11.2.0.1 AWM that can cause the symptoms you are having. It may be related to the following AWM bug.
BUG 8889693 - SUMMARY LEVEL MAPPING DOES NOT WORK IF USING DIMENSIONS WITH SURROGATE KEYS
There is an AWM "A Patch" for 11.2.0.1 (Patch 9672534: AWM 11.2.0.1A) that fixed the above bug and several other mapping related problems. You could try this to see if it fixes your problem.

Cube Processing approach when Process only the Current Partition?

Could you validate my SSAS Processing strategy for the given scenarios:<o:p></o:p>
Background about cube and data:<o:p></o:p>
A Sales cube has Partitions for each year for "Sales" Measure Group and it associated with dimensions "Product" and "Sales
Rep". Both are type 1 Dimensions.<o:p></o:p>
Here some time user will re-classify the products, product Hierarchy(Product -> Sub Category - > Category); <o:p></o:p>
Similarly re-classify Sales Rep (Sales Rep –> District Manager -> Regional Manager)<o:p></o:p>
Processing strategy:<o:p></o:p>
1. Process(full process) only the current Partition every day.<o:p></o:p>
2. Perform "Process update” for all the dimensions. (going for process update, as Dimension full process is processing all the old partitions
of measure groups)<o:p></o:p>
Questions:<o:p></o:p>
1.
What are disadvantages when processing only the current partition? <o:p></o:p>
2.
Does the old partitions data will roll up as per the hierarchy changes, when I go for Dimension “process update” options.<o:p></o:p>
Thanks,
Liyasker Samraj K

1.
What are disadvantages when processing only the current partition? <o:p></o:p>
2.
Does the old partitions data will roll up as per the hierarchy changes, when I go for Dimension “process update” options.<o:p></o:p>
The strategy looks good. Partitioning is the way to go to reduce processing time. However, keep in mind that partitions are only supported in the enterprise version.
1. Other than not being able to refresh older data from other partitions, i don't see a downside in processing the most recent partitions.
2. Yes. A process update should touch all the dependent partitions.
SS

Cube partitioning

Hi Gurus,
I have more than 2 lakh record in my infocube , now to optimize the performance i read about cube partitioning, can anyone explain me how to do cube partitioning , steps and on what basic we do cube partitioning.
Thanks & Regards,
Lakshmi Rajkumar.

Hi Lakshmi,
You can only partition on 0FISCPER or 0CALMONTH InfoObject. For Partioning you need to maintain atleast one of the mentioned InfoObject in your InfoProvider.
For more info. please check the below help file:
http://help.sap.com/saphelp_nw70/helpdata/en/33/dc2038aa3bcd23e10000009b38f8cf/frameset.htm
Hope it helps.
Regards,
Raghu

HR cube for employee and total gross compensation

Hi,
I want to make a query which contains total number of employee and total gross compensation.
can anyone suggest me any cube for this? ı am new on BW and ı cant give a decision about cubes.
thanks,

Hi,
Have a look at the below link.
http://help.sap.com/saphelp_nw70/helpdata/en/3c/f3e8372ad84a13e10000009b38f8cf/frameset.htm
Regards,
Sunil

ORA-36836 about cube dimensionality ???

Similar Messages

Maybe you are looking for