Aggregate/ Multicube

BW Experts,
I am novice to BW! Hence I these basic questions-
1. If data is deleted from the cube what happens to the data in the aggregate built on that cube. WIll it be deleted automatically?
2. Can aggregates be built on the multicube ?
Well the solution I have been asked to provide by the client is
Scenario: We have a cube that holds data all the way from 2001. There is a aggregate built on this cube.
Now I need to split data into 5 different cubes according to the year say 2001, 2002, 2003, 2004,2005.
What i have figured to do is to split data into different cubes and build a multicube on these 5 cubes. Also once the data is split into 5 cubes it needs to be deleted from the parent cube and the parent cube will hold only data pertaining to 2006. I guess you all understand my question better now...So when i delete data from 2001-2005 from the parent cube, what happens to the data in the aggregate built on the parent cube?
Well I am not sure whether the solution I came up with is the best one. Correct me if I am wrong.
Please help this new comer!!!
Thanks
Ashwin

You have two choices - you can run a selective deletion on the cube for just 2001, or you could use the archive function to archive off 2001 data. Same net result, system just takes two different approaches - depends on your need to retain the data somewhere and familiarity with either function.
The selective deletion process actually checks to see if your deletion criteria would result in deleting more than 10% of the data in cube. If it will (and it sounds like it will in your case) the BW actually creates a copy of the cube and copies all the data except for what you want to delete, to the new cube, then deletes the old cube and renames the new cube. It does this because as mentioned earlier, DB deletes are resource intensive transactions, and the DB can actually copy 90% of the cube to a new cube faster than it can delete 10% of the data form the original cube.
So one thing you need to keep in mind any time you plan to run a selective deletion that might delete more than 10% of the data, you need to verify that your Infocube fact table tablespaces have sufficent room to temporarily hold the original cube and the copy.
A rather long - "yes" answer, isn't it?

Similar Messages

Aggregates not in use

Hi,
How to get the list of the aggregates which are currently not in use, without going to the aggregates maintainence screen?
Thanks

Samruddhi,
Did you try in LISTSCHEMA?
Also check this useful thread
Re: Aggregate/ Multicube
Hope it helps,
*******Assign Points******
Thanks
Gattu

DBMAN350 Error reading MultiCube data over aggregate

Hi, I have handed down multicube over a Cube, DSO, and IP Aggregate. When I try to display the data using a LISTCUBE I get errors. I just noticed I get the same error in my development box.
1. Error reading the data of InfoProvider Z_C017_A8 Message no. DBMAN305
Diagnosis - Errors occurred while reading a VirtualProvider outside the BI system. Check whether the previous error messages contain any information about the possible cause of this error.
It is possible that the error message cannot be displayed because the error message class does not exist in the BI system. If this is the case, only the name of the error class and the message number are displayed. View the error message text in the specified error class in the source system of the VirtualProvider.
Procedure - Since the error is not necessarily in the BI system, there is no specific procedure for resolving it. With VirtualProviders, problems often occur with the connection to the remote system; these can lead to system termination. If the code for the VirtualProvider is not from SAP, contact the relevant contact person to help resolve the issue.
If an SQL error is listed in the previous message, see the procedure for SQL errors.
2. Reading from InfoProvider Z_C017_A8 is not permitted Message no. RSPLS801
Diagnosis - Cannot read data from InfoProvider Z_C017_A8 here.
System Response - The system has cancelled the read operation.
Procedure - Use InfoProvider Z_P_C017 for the read operation here
I am suspecting this is why data from the planning aggregate is not showing up on my query, but the query doesn't return an error like listcube does. It is OK to have planning aggregates in a multiprovider I am assuming because it allows it.
Any ideas why this error is appearing?
I already searched SDN and OSS, but nothing relevant appears.

Hi Kenneth,
I found this notes which could be relevant for you:
Error when reading RemoteCubes in MultiProviders
SAP Note Number: [1011321|https://www.sdn.sap.com/irj/servlet/prt/portal/prtroot/com.sap.km.cm.docs/oss_notes/sdn_oss_bw_bex/~form/handler]
DBMAN 305: Error when reading data of InfoProvider ...$N
SAP Note Number: [1051291|https://www.sdn.sap.com/irj/servlet/prt/portal/prtroot/com.sap.km.cm.docs/oss_notes/sdn_oss_bw_whm/~form/handler]
Additional this thread:
Remote cube - uncaught exception
Regards
Andreas

Help on Creating Aggregates on MultiProvider

Hi All,
Is it possible to create aggregates on MultiProvider or its possible in Basic Cubes Only. please suggest with that? some document i go through You do not have to create aggregates for MultiCubes.
regards
babu

magesh,
aggregates on multiprovider would mean separate aggregates on the constituent cubes / ODS...
Go to the cube and select propose aggregate from query and select your multiprovider query - use the aggregate(s) proposed for the same....
When running the query on multiprovider , the agregates on the constituent cubes will be used...
Arun
Hope it helps...

How to create suitable aggregates for queries on multiprovider ?

hi all,
goal reduce db time of query and improve performance
i have queries on a multicube. I have 5 cubes under the multiprovider. I having performance issue with one of the cubes. it had
high slection/transfer ratio. The same cube had high 94% DB TIme. All the BW and DB indexes and stats are green. I chose
the path of aggregates. when i tried suggest from proposal it is giving me query time and date range and i gave last
3 days and query time 150 sec. it is suggesting huge number of aggregates like 150 of them and not getting reduced much
when i tried optimize funcitonality.
The faulty cube had nearly 9 million records and 4 years of data
1. generally how many aggregates do we need to create on a cube?
2. how do i use propose from last navigation? it is not creating any aggregates
3. is there a way for system to propose less number of aggregates?
4. if nothing works i want to cut the volume of the aggregates base on years or quarters. how do i do that?
i created with time charactersitic 0calquarter and dragged in ocal day and 0calmonth. activated and filled in ..but
query is not hitting it when i do a monthly selection. i tried bringing in all the other dimensions...except line item
dimenisons....no use ...it is not hitting the manual aggregates in RSRT.the slection on 0calquarter is * .
5. should i change it to fixed value and bring in the line items too and create it?
6. I wanted to try propose aggregate from query option..but my query is on multiprovider...and not able to copy it to cube..
plz help me how to find the suitable aggregates for query on multiprovider
7. should i create any new indexes on the cube using the chractrestics in the where condiotion of select statment...but in
that case, select statement changes with drill down.......how do i handle it ?
8. how will i make sure the aggregates work for all queries run time imporvement?
9. plz suggest other approaches if any with procedures
this is a urgent problem plz help...
<b>thanks in advance
points will be assigned for inputs</b>

1. generally how many aggregates do we need to create on a cube?
it depends on your specific needs, you can need none or several.
2. how do i use propose from last navigation? it is not creating any aggregates
Can you elaborate?
3. is there a way for system to propose less number of aggregates?
In any of the menus of screen for creating aggregates you have an option for sytem propose aggregates for one specific queries i am not sure it worked with multicubes.
4. if nothing works i want to cut the volume of the aggregates base on years or quarters. how do i do that?
You should delete 0calday from aggregates in order to accumulate data for any greater time unit. Other solution for times is to study try to do a partition in cube.
5. should i change it to fixed value and bring in the line items too and create it?
Can you elaborate?
6. I wanted to try propose aggregate from query option..but my query is on multiprovider...and not able to copy it to cube..
Answered before, maybe you can create a query only with data on that cube that appears in multicube query in order to proposal any aggregate in thath cube.
7. should i create any new indexes on the cube using the chractrestics in the where condiotion of select statment...but in
that case, select statement changes with drill down.......how do i handle it ?
is not recomendable create new indexes in multidemensional structures. Try avoid selections for navigational attributes, if necesary add navigate attributte as dimension attributes, put filters in filter section in BEX.
8. how will i make sure the aggregates work for all queries run time imporvement?
try transaction st03
9. plz suggest other approaches if any with procedures
Some other approches yet answering
Good luck

With new Free Characeristics Aggregates are not being used

Hi....i have a query from a multicube with two infocubes where i have two aggregates one for each cube...in my query both aggregates are being used with no problems....when i add a new free characteristic that appers on aggregates something happens and the aggregates are not being used anymore...i dont understand why this happens because i know that ......free characteristics ....doesnt affects the aggregates usage....is this correct??? or not?
Regards

Since this is an open forum, not Adobe support... you need to contact Adobe staff to help
Adobe contact information - http://helpx.adobe.com/contact.html
-Select your product and what you need help with
-Click on the blue box "Still need help? Contact us"
-or by telephone http://helpx.adobe.com/x-productkb/global/phone-support-orders.html

Problem: Aggregates for repository by scheduling an aggregate creation job

Hi dear OBIEE Gurus,
I‘m trying to create my Aggregates for my repository by scheduling an aggregate creation job as shown in
http://www.oracle.com/technology/obe/obe_bi/bi_ee_1013/aggpersist/aggpersist.htm
env:
1-     Oracle Business Intelligence Product Version 10.1.3.4.1 (Build 090414.1900)
2-     Windows XP
3-     Three separate repositories (OE, SH, & Sample Sales) running in parallel each on its own analytics deployment on one OC4J server
+++++++++++++++++
What is the problem:
When I try to start the scheduled job I get following message:
+++++++++++++++++
Oracle BI Server
Copyright (c) 1997-2009 Oracle Corporation, All rights reserved
delete aggregates
delete aggregates
Statement execute succeeded
create aggregates
"ag_Salesfacts"
for "SH"."Salesfacts"("Amount Sold")
at levels ("SH"."ProductsDim"."Category", "SH"."TimesDim"."Times Detail", "SH"."CustomersDim"."State Province")
using connection pool "orcl SH"."Connection Pool"
in "orcl SH".."SH"
create aggregates
"ag_Salesfacts"
for "SH"."Salesfacts"("Amount Sold")
at levels ("SH"."ProductsDim"."Category", "SH"."TimesDim"."Times Detail", "SH"."CustomersDim"."State Province")
using connection pool "orcl SH"."Connection Pool"
in "orcl SH".."SH"
Processed: 2 queries
Encountered 1 errors
[NQODBC] [SQL_STATE: HY000] [nQSError: 10058] A general error has occurred.
[nQSError: 46036] Internal Assertion: Condition !m_factTable.IsNull(), file .\Src\SAAPAggrCube.cpp, line 46.
Statement preparation failed
++++++++++++++++++++
What have I done so far?
++++++++++++++++++++
-     I went through all steps in the aforementioned tutorial, up to the point where it says “6. Verify that the job completes successfully in the Job Manager.”
-     I set up the job in the scheduler
-     I let run the job in the job manager, Which is where I get the error.
-     I tried different combinations of fields, aggregations and levels, all with the same result.
-     Manually starting the job has the same result: >D:\programs\oracle\obiee\server\Bin\NQCmd.exe -u Administrator -p Administrator -d AnalyticsWebSH -s D:\programs\oracle\obiee\server\Repository\create_aggregated_table_rk.SQL
-     I tried to create aggregates in my other repositories, and found OTHER errors, but at least the aggregates were created there (to some extent)
-     I looked in OTN for the error, found nothing
-     Google does not know a thing about this
Any ideas?

Hi John,
there is an OSS Note 453026 'No aggregate proposals for MultiCubes after query' that seems explaining your situation, but I think you aren't on the old 2.0/2.1 versions...
John, I don't know how could I help you...
If I find more, I'll write you at once !
Bye,
Roberto

About Multicube vs a giant cube with Aggr

Hi Gurus,
i got this question, whay is the best decition about design,
1- create a cube per year that is feed from ECC , for example cube data 2011 and cube data2012 and above this one, a Multicube, then the report
2- Create a Giant Cube with aggregates
help guys,

Hi Jorge
I would go for option one but still have aggregates.
Option 1 actually splits your data in different cubes. Cubes with less data is always favorable for administrative work and report runtime.
Like compression, roll up, creation and deletion of Index is easier in small cubes. Moreover if there is any issue in data level you can always check from smaller cubes fast. If the cube is very large Display Data is a pain.
One more point is , once you load historical data in different cubes, you do not need to touch them because they are already frozen. You just need to take care of the cube which contains current data.
Aggregates on the other hand takes extra space on the database. Administration of aggregates is also required . Like daily roll up adds to your data load time. More over aggregates are very specific to query. One aggregate can not satisfy all the query.
It's better to have different cubes and then you can have small aggregates on those cubes. This would make your administration easy and query response time will be faster.
PS: If multiprovider has different cubes for 2010,2011 and 2012 even if the query only wants 2012 data , backend system would read 2010 and 2011 cubes also. You can have a filter on 0INFOPROV in query level to prevent that.
Regards
Anindya

MultiCube over IP Aggregation Level Question

Hi, I have a MultiProvider over an IP Aggregation Level and a Real InfoCube. This was all handed down to me from previous development. The Aggregatation Level has the same level of detail as the planning infocube it is based on.
Is there any reason for the MultiProvider to be over the Aggregation Level instead of the actual planning infocube?
I guess I'm reading that a multiprovider over an aggregation level is for planning only and not reporting? If this is true then do I need a multicube for planning and a separate multicube for reporting using the actual planning infocube?
Thanks for any clarity you can offer...

Wouldn't this modeling work as well for planning? I mean while they are planning values they might want to see actual values as well to help in decision making?
Thanks!

Index's on cubes or Aggregates on infoobjects

Hello,
Please tell me if it is possible to put index's on cubes; are they automatically added or is this something I put on them?
I do not understand index's are they like aggregates?
Need to find info that explains this.
Thanks for the hlep.
Newbie

Indexes are quite different from aggregates.
An Aggregate is a slice of a cube which helps the data retrival on a faster note when a query is executed on a cube. Basically it is kind of a snapshot of KPI's and Business Indicators (Chars) which will be displayed as the initial query run result.
Index is a process which is inturn will reduce the query response time. While an object gets activated, the system automatically create primary indexes. Optionaly, you can create additional index called secondary indexes.Before loading data, it is advisable to delete the indexes and insert them back after the loading.
Indexes act like pointers for quickly geting the Data.When u delete it will delete indexes and when u create it will create the indexes.
When loading we delete Bcs during loading it has to look for existing Indexes and try to update so it will effect the Data load performence so we delete and create it will take less time when compared to updating the existing ones.
There is one more issue we have to take care if u r having more than 50 million records this is not a good practice insteah we can delete and create during week end when they r no users.

How can I see the data in the aggregates

how can see the data available in the aggregates.
Jay

Hi Jay,
its so simple,
please goto the manage aggregates screen and copy the technical name of the aggregate and add
/bic/exxx xxx is the aggregate technical name, and for f fat table use /bic/fxxx, and go to se16 and enter the table name and thats it ur data is with u.
R

Aggregates, VLAN's, Jumbo-Frames and cluster interconnect opinions

Hi All,
I'm reviewing my options for a new cluster configuration and would like the opinions of people with more expertise than myself out there.
What I have in mind as follows:
2 x X4170 servers with 8 x NIC's in each.
On each 4170 I was going to configure 2 aggregates with 3 nics in each aggregate as follows
igb0 device in aggr1
igb1 device in aggr1
igb2 device in aggr1
igb3 stand-alone device for iSCSI network
e1000g0 device in aggr2
e1000g1 device in aggr2
e1000g2 device in aggr3
e1000g3 stand-alone device of iSCSI network
Now, on top of these aggregates, I was planning on creating VLAN interfaces which will allow me to connect to our two "public" network segments and for the cluster heartbeat network.
I was then going to configure the vlan's in an IPMP group for failover. I know there are some questions around that configuration in the sense that IPMP will not detect a nic failure if a NIC goes offline in the aggregate, but I could monitor that in a different manner.
At this point, my questions are:
[1] Are vlan's, on top of aggregates, supported withing Solaris Cluster? I've not seen anything in the documentation to mention that it is, or is not for that matter. I see that vlan's are supported, inluding support for cluster interconnects over vlan's.
Now with the standalone interface I want to enable jumbo frames, but I've noticed that the igb.conf file has a global setting for all nic ports, whereas I can enable it for a single nic port in the e1000g.conf kernel driver. My questions are as follows:
[2] What is the general feeling with mixing mtu sizes on the same lan/vlan? Ive seen some comments that this is not a good idea, and some say that it doesnt cause a problem.
[3] If the underlying nic, igb0-2 (aggr1) for example, has 9k mtu enabled, I can force the mtu size (1500) for "normal" networks on the vlan interfaces pointing to my "public" network and cluster interconnect vlan. Does anyone have experience of this causing any issues?
Thanks in advance for all comments/suggestions.

For 1) the question is really "Do I need to enable Jumbo Frames if I don't want to use them (neither public nore private network)" - the answer is no.
For 2) each cluster needs to have its own seperate set of VLANs.
Greets
Thorsten

Aggregates on Non-cumulative InfoCubes, stock key figures, stock, stocks,

Hi..Guru's
Please let me know if anybody has created aggregates on Non-Cumulative Cubes or key figure (i.e. 0IC_C03 Inventory Management.)
I am facing the problem of performance related at the time of execution of query in 0IC_C03.( runtime dump )
I have tried lot on to create aggregate by using proposal from query and other options. But its not working or using that aggr by query.
Can somebody tell me about any sample aggr. which they are using on 0ic_c03.
Or any tool to get better performance to execute query of the said cube.
One more clarification req that what is Move the Marker pointer for stock calculation. I have compressed only two inital data loading req. should I compress the all req in cube (Regularly)
If so there would be any option to get req compress automatically after successfully load in data target.
We are using all three data sources 2lis_03_bx,bf & um for the same.
Regards,
Navin

Hi,
Definately the compression has lot of effect on the quey execution time for Inventory cubes <b>than</b> other cumulated cubes.
So Do compression reqularly, once you feel that the deletion of request is not needed any more.
And ,If the query do not has calday characterstic and need only month characterstic ,use Snap shot Info cube(which is mentioned and procedure is given in How to paper) and divert the month wise(and higher granularity on time characterstic ,like quarter & year) queries to this cube.
And, the percentage of improvement in qury execution time in case of aggregates is less for non cumulated cubes when compared to other normal(cumulated) cubes. But still there is improvement in using aggregates.
With rgds,
Anil Kumar Sharma .P
Message was edited by: Anil Kumar Sharma

Re: How to Improve the performance on Rollup of Aggregates for PCA Infocube

Hi BW Guru's,
I have unresolved issue and our team is still working on it.
I have already posted several questions on this but not clear on how to reduce the time on Rollup of Aggregates process.
I have requested for OSS note and searching myself but still could not found.
Finally i have executed one of the cube in RSRV with the database selection
"Database indexes of an InfoCube and its aggregates" and got warning messages i was tried to correct the error and executed once again but still i found warning message. and the error message are as follows: (this is only for one info cube we got 6 info cubes i am executing one by one).
ORACLE: Index /BI0/IACCOUNT~0 has possibly degenerated
ORACLE: Index /BI0/IPROFIT_CTR~0 has possibly degenerated
ORACLE: Index /BI0/SREQUID~0 has possibly degenerated
ORACLE: Index /BIC/D1001072~010 has possibly degenerated
ORACLE: Index /BIC/D1001132~010 has possibly degenerated
ORACLE: Index /BIC/D1001212~010 has possibly degenerated
ORACLE: Index /BIC/DGPCOGC062~01 has possibly degenerated
ORACLE: Index /BIC/IGGRA_CODE~0 has possibly degenerated
ORACLE: Index /BIC/QGMAPGP1~0 has possibly degenerated
ORACLE: Index /BIC/QGMAPPC2~0 has possibly degenerated
ORACLE: Index /BIC/SGMAPGP1~0 has possibly degenerated
i don't know how to move further on this can any one tell me how to tackle this problem to increase the performance on Rollup of Aggregates (PCA Info cubes).
every time i use to create index and statistics regularly to improve the performance it will work for couple of days and again the performance of the rollup of aggregates come down gradually.
Thanks and Regards,
Venkat

hi,
check in a sql client the sql created by Bi and the query that you use directy from your physical layer...
The time between these 2 must be 2-3 seconds,otherwise you have problems.(these seconds are for scripts that needed by Bi)
If you use "like" in your sql then forget indexes....
For more informations about indexes check google or your Dba .
Last, i mentioned that materialize view is not perfect,it help a lot..so why not try to split it to smaller ones....
ex...
logiacal dimensions
year-half-day
company-department
fact
quantity
instead of making one...make 3,
year - department - quantity
half - department - quantity
day - department - quantity
and add them as datasource and assign them the appropriate logical level at bussiness layer in administrator...
Do you use partioning functionality???
i hope i helped....
http://greekoraclebi.blogspot.com/
///////////////////////////////////////

Questions regarding aggregates on cubes

Can someone please answer the following questions.
1. How do I check whether someone is re-bilding aggregates on a cube?
2. Does rebuilding an aggregate refer to the rollup process? Can it take a few hours?
3. What does it mean when someone switches off an aggregate, basically what is the difference (conceptually/time consumption)between:
                        A. activating an aggregate?
                        B. switching off/on an aggregate?
                        C. rebuilding an aggregate?
3. When a user complains that a query is running slow, do we build an aggregate based on the chars in rows & free chars in that query OR is there anything else we need to include?
4. Does database statistics in the 'MANAGE' tab of a cube only show statistics or does it do anything to improve the load/query performance on the cube?
Regards,
Srinivas.

1. How do I check whether someone is re-bilding aggregates on a cube?
If your aggregate status is in red and you are filling up the aggregate - it is an initial fill of the aggregate and filling up would mean loading the data from the cube into the aggregate in full.
2. Does rebuilding an aggregate refer to the rollup process? Can it take a few hours?
Rebuilding of an aggregate is to reload the data into the aggregate from the cube once again.
3. What does it mean when someone switches off an aggregate, basically what is the difference (conceptually/time consumption)between:
A. activating an aggregate?
this would mean recreating the data structures for the aggregate - this would mean dropping the data and reloading the data.
B. switching off/on an aggregate?
Switching off an aggregate means that it will not be used by the OLAp processor but would mean that the aggregate still gets rolled up. Rollup referring to loading changed data from the cube into the aggregate - this is done based n the requests that have not yet been rolled up into the cube.
C. rebuilding an aggregate?
Reloading data into the aggregate
3. When a user complains that a query is running slow, do we build an aggregate based on the chars in rows & free chars in that query OR is there anything else we need to include?
Run the query in RSRT and do an SQl view of the query and check the characteristics that are used in the query and then include the same into your aggregate.
4. Does database statistics in the 'MANAGE' tab of a cube only show statistics or does it do anything to improve the load/query performance on the cube?
Stats being updated will improve the execution plans on the database. Making sure that stats are up to date would lead to better execution plans and hence possibly better performance but it cannot eb taken for granted that refreshing stats is going to improve query performance.

Aggregate/ Multicube

Similar Messages

Maybe you are looking for