Sunday, February 27, 2011

Partitioning on W_GL_OTHER_F - 1

In our database, W_GL_OTHER_F table has more then 70 million records, and reports have already started hitting bad performance. Considering we will go live on more countries in our global OBIEE project, the number of records is getting higher day by day. According to Oracle Business Intelligence Applications Version 7.9.6 Performance Recommendations (BI_Apps796_Perf_Tech_Note_V5), we have decided to go on partitiopning over W_GL_OTHER_F table. I will try to write the all steps that we followed to help other people who are trying to do the same, as we had many questions before starting this process, and we could not find quick answers by just googling.

We are implementing the range partitioning with ACCT_PERIOD_END_DT_WID and also partitioning the data by quarterly. The first problem that we hit was how to partition the data by quarterly while the ACCT_PERIOD_END_DT_WID is a number field having populated by concatenating 3 values
M_CAL_CAL_WID || ACCT_PERIOD_END_DT|| IIF(ADJUSTMENT_FLG='Y', '999', '000')
M_CAL_CAL_WID is an ID of calendar being used, in our case it is 1003. And the last 3 digits has not much impact on partitioning. Like if accounting period end is 20101031 (31 Oct 2010) and if it is adjustment record then ACCT_PERIOD_END_DT_WID would be 100320101031999.  Therefore we came up with the solution that we will be loading this data into 4th Quarter of 2010, which can be defined as "partition PART_2010Q4 values less than (100320101232000)", see the below create table script for more detail, but itis simply that we do not need to comply with date formatting, it is a number in this column anyway, so we can play with it as we want.

Steps to be done in Datawarehouse Database
Rename the original table
rename W_GL_OTHER_F to W_GL_OTHER_F_ORIG;
  
Before running the create table ddl, execute the below query in database and change the partition range value accordingly.
SELECT row_wid
FROM w_mcal_cal_d
WHERE mcal_cal_name = 'YOUR CALENDAR'
Please use the value from the above query while defining the partition range. As in our environment the value is 1003 we have used 1003**** and replace this with the query result.

Create the partitioned table, using range partitioning by quarter;
Create table W_GL_OTHER_F partition by range (ACCT_PERIOD_END_DT_WID)
(
partition PART_MIN values less than (100320061232000),
partition PART_2007Q1 values less than (100320070332000),
partition PART_2007Q2 values less than (100320070632000),
partition PART_2007Q3 values less than (100320070932000),
partition PART_2007Q4 values less than (100320071232000),
partition PART_2008Q1 values less than (100320080332000),
partition PART_2008Q2 values less than (100320080632000),
partition PART_2008Q3 values less than (100320080932000),
partition PART_2008Q4 values less than (100320081232000),
partition PART_2009Q1 values less than (100320090332000),
partition PART_2009Q2 values less than (100320090632000),
partition PART_2009Q3 values less than (100320090932000),
partition PART_2009Q4 values less than (100320091232000),
partition PART_2010Q1 values less than (100320100332000),
partition PART_2010Q2 values less than (100320100632000),
partition PART_2010Q3 values less than (100320100932000),
partition PART_2010Q4 values less than (100320101232000),
partition PART_2011Q1 values less than (100320110332000),
partition PART_2011Q2 values less than (100320110632000),
partition PART_2011Q3 values less than (100320110932000),
partition PART_2011Q4 values less than (100320111232000),
partition PART_2012Q1 values less than (100320120332000),
partition PART_2012Q2 values less than (100320120632000),
partition PART_2012Q3 values less than (100320120932000),
partition PART_2012Q4 values less than (100320121232000),
partition PART_MAX values less than (maxvalue)
)
TABLESPACE XXXDW_DATA
PCTUSED    0
PCTFREE    10
INITRANS   1
MAXTRANS   255
STORAGE    (
            INITIAL          4M
            NEXT             4M
            MINEXTENTS       1
            MAXEXTENTS       2147483645
            PCTINCREASE      0
            BUFFER_POOL      DEFAULT
           )
nologging parallel
enable row movement
as (select * from W_GL_OTHER_F_ORIG);
/
Rename indexes on renamed table
ALTER INDEX W_GL_OTHER_F_71 rename to W_GL_OTHER_F_71_ORIG;
ALTER INDEX W_GL_OTHER_F_72 rename to W_GL_OTHER_F_72_ORIG;
ALTER INDEX W_GL_OTHER_F_73 rename to W_GL_OTHER_F_73_ORIG;
ALTER INDEX W_GL_OTHER_F_F59 rename to W_GL_OTHER_F_F59_ORIG;
ALTER INDEX W_GL_OTHER_F_F25 rename to W_GL_OTHER_F_F25_ORIG;
ALTER INDEX W_GL_OTHER_F_F36 rename to W_GL_OTHER_F_F36_ORIG;
ALTER INDEX W_GL_OTHER_F_F40 rename to W_GL_OTHER_F_F40_ORIG;
ALTER INDEX W_GL_OTHER_F_F5 rename to W_GL_OTHER_F_F5_ORIG;
ALTER INDEX W_GL_OTHER_F_M2 rename to W_GL_OTHER_F_M2_ORIG;
ALTER INDEX W_GL_OTHER_F_C70 rename to W_GL_OTHER_F_C70_ORIG;
ALTER INDEX W_GL_OTHER_F_F13 rename to W_GL_OTHER_F_F13_ORIG;
ALTER INDEX W_GL_OTHER_F_F11 rename to W_GL_OTHER_F_F11_ORIG;
ALTER INDEX W_GL_OTHER_F_F60 rename to W_GL_OTHER_F_F60_ORIG;
ALTER INDEX W_GL_OTHER_F_F61 rename to W_GL_OTHER_F_F61_ORIG;
ALTER INDEX W_GL_OTHER_F_U1 rename to W_GL_OTHER_F_U1_ORIG;
Create Global and Local indexes. Run the below statements to create indexes
CREATE BITMAP INDEX W_GL_OTHER_F_C70 ON W_GL_OTHER_F (LOC_CURR_CODE ASC) tablespace XXXDW_DATA LOCAL NOLOGGING;
CREATE BITMAP INDEX W_GL_OTHER_F_72 ON W_GL_OTHER_F (X_REFERENCE_8 ASC) tablespace XXXDW_DATA LOCAL NOLOGGING;
CREATE UNIQUE INDEX W_GL_OTHER_F_U1 ON W_GL_OTHER_F (INTEGRATION_ID ASC,DATASOURCE_NUM_ID ASC) tablespace XXXDW_DATA GLOBAL NOLOGGING;
CREATE BITMAP INDEX W_GL_OTHER_F_73 ON W_GL_OTHER_F (DOC_CURR_CODE ASC) tablespace XXXDW_DATA LOCAL NOLOGGING;
CREATE BITMAP INDEX W_GL_OTHER_F_71 ON W_GL_OTHER_F (DOC_STATUS_WID ASC) tablespace XXXDW_DATA LOCAL NOLOGGING;
CREATE BITMAP INDEX W_GL_OTHER_F_F11 ON W_GL_OTHER_F (COMPANY_ORG_WID ASC) tablespace XXXDW_DATA LOCAL NOLOGGING;
CREATE BITMAP INDEX W_GL_OTHER_F_F13 ON W_GL_OTHER_F (COST_CENTER_WID ASC) tablespace XXXDW_DATA LOCAL NOLOGGING;
CREATE BITMAP INDEX W_GL_OTHER_F_F25 ON W_GL_OTHER_F (GL_ACCOUNT_WID ASC) tablespace XXXDW_DATA LOCAL NOLOGGING;
CREATE BITMAP INDEX W_GL_OTHER_F_F36 ON W_GL_OTHER_F (POSTED_ON_DT_WID ASC) tablespace XXXDW_DATA LOCAL NOLOGGING;
CREATE BITMAP INDEX W_GL_OTHER_F_F40 ON W_GL_OTHER_F (PROFIT_CENTER_WID ASC) tablespace XXXDW_DATA LOCAL NOLOGGING;
CREATE BITMAP INDEX W_GL_OTHER_F_F5 ON W_GL_OTHER_F (BUSN_AREA_ORG_WID ASC) tablespace XXXDW_DATA LOCAL NOLOGGING;
CREATE INDEX W_GL_OTHER_F_F59 ON W_GL_OTHER_F (ACCT_PERIOD_END_DT_WID ASC) tablespace XXXDW_DATA LOCAL NOLOGGING;
CREATE BITMAP INDEX W_GL_OTHER_F_F60 ON W_GL_OTHER_F (LEDGER_WID ASC) tablespace XXXDW_DATA LOCAL NOLOGGING;
CREATE BITMAP INDEX W_GL_OTHER_F_F61 ON W_GL_OTHER_F (MCAL_CAL_WID ASC) tablespace XXXDW_DATA LOCAL NOLOGGING;
CREATE BITMAP INDEX W_GL_OTHER_F_M2 ON W_GL_OTHER_F (DELETE_FLG ASC) tablespace XXXDW_DATA LOCAL NOLOGGING;
And finally compute statistics on partitioned table
BEGIN
dbms_stats.Gather_table_stats(
NULL,
tabname => 'W_GL_OTHER_F',
CASCADE => true,
estimate_percent => dbms_stats.auto_sample_size,
method_opt => 'FOR ALL INDEXED COLUMNS SIZE AUTO');
END;
Database part is done, we need to do DAC changes that I will try to post later. Actually for DAC part, the document from Oracle is more than enough, but I will repeat some steps to finalize the document. My team from offshore has done good job on delivering this solution, a big thanks to them.

Thursday, February 10, 2011

Vertical Text Title in OBIEE Answers

We had a requirement that report needs to show column titles vertically in the Answers report. We used to have similar thing for downloaded excel reports by using "mso-rotate:90" in "Use Custom CSS Style" before. But this time requirement is for answer reports. Similarly by adding the following statement into column heading formatting part, we achieved to have vertically aligned heading in the report as seen below;

"{writing-mode: tb-rl;filter: flipv fliph;}"

 

 

Saturday, January 29, 2011

Override Session Variable in Answers

If you wanted to override a session variable and disable cache hits, you would use a syntax like the following in Answers Advanced Tab
SET VARIABLE LOGLEVEL=4, DISABLE_CACHE_HIT = 1;

SET VARIABLE OU_ORG=128;

 

Tuesday, January 25, 2011

Performance Tuning Tips for OBIEE


1. implement caching mechanism
2. use aggregates
3. use aggregate navigation
4. limit the number of initialisation blocks
5. turn off logging
6. carry out calculations in database
7. use materialized views if possible
8. use database hints
9. alter the NQSONFIG.ini parameters

http://www.oraclebidwh.com/2010/02/performance-tuning-in-obiee/

Sunday, January 23, 2011

OBIEE is not an ETL tool!

just a reminder for whom using OBIEE as Excel Extraction tool;

"OBIEE is not an ETL tool. It is NOT designed to handle large amounts of non aggregated data. It’s designed to handle dimensional structured data with sufficient aggregation tables."

Friday, January 21, 2011

Changing the Connection Pool from Answers



Tuesday, January 04, 2011

Keeping Leading Zero's in Answers

Nice way to keep leading zero’s when exporting to excel from within OBIEE:

Add “mso-number-format:\@” to the Custom CSS Style Options of the appropriate column in the OBIEE report.

Excel rotating style

mso-rotate:90

Saturday, December 04, 2010

UDDI Connection in JDeveloper

 

In the Connection Wizard, provide a connection name and specify the UDDI inquiry endpoint URL. The syntax of this URL is:

http://ohs_host:ohs_Port/registry_context/uddi/inquiry

ohs_host and ohs_Port have the following definitions:

·         ohs_host is the address of the Oracle Application Server host machine.; for example, server07.company.com

·         ohs_Port is the HTTP listener port assigned to OHS

registry_context is context root used to access the target registry instance, such as "registry" or "registrypub"

Tuesday, November 02, 2010

How to export Security Group to create UDML


Here is a command to extract security objects from an OBIEE repository using the UDML command, mind the “-S” in the end, it handles to export Security part only;

C:\oracle\OBIEE\OracleBI\server\Bin\nQUDMLGen.exe -U Administrator -P xxx123 -R D:\ALL\repository\temps\yaz\050310_XXX_BAW_CS.rpd -O D:\ALL\repository\temps\yaz\tekin.udml -S

C:\OracleBI\server\bin\nqudmlgen -U Administrator -P Administrator -R C:\OracleBI\server\Sample\samplesales\samplesales.rpd -O C:\OracleBI\server\Sample\samplesales\samplesales.udml -S

nqudmlgen: This is the actual export command
-U: This is the user flag, in my case Administrator
-P: This is the password flag, in my case Administrator since I'm using sample sales
-R: Is the source repository. I am sourcing from samplesales.rpd on a Windows box
-O: Is the output file. I am populating a file called samplesales.udml in the same folder as the source
-S: Exports only security objects

Here is the command to import the objects

nqudmlexec -U Administrator -P Administrator -I C:\OracleBI\server\Sample\samplesales\samplesales.udml -B C:\OracleBI\server\Sample\samplesales\samplesales.rpd -O C:\OracleBI\server\Sample\samplesales\samplesales2.rpd

nqudmlexec : This is the actual import command
-U: This is the user flag, in my case Administrator
-P: This is the password flag, in my case Administrator since I'm using sample sales
-I: This is the input script. In this case, the input to this is the output from the first command
-B: This is the base repository. In this case, your base repository will be your target repository, the one where you want to migrate the users to.
-O: This is the output repository. This command makes a copy and applies the changes in the UDML to that copy. So you need to specify an output file. This is the file that will have the changes. The base file will not have the new users since it is just used to make this copy.
 Example;


EXPORT;
C:\oracle\OBIEE\OracleBI\server\Bin\nQUDMLGen.exe -U Administrator -P xxx-R D:\ALL\repository\temps\yaz\050310_XXX_BAW_CS.rpd -O D:\ALL\repository\temps\yaz\tekin.udml

IMPORT;
C:\oracle\OBIEE\OracleBI\server\Bin\nQUDMLexec.exe -U Administrator -P xxx -I D:\ALL\repository\temps\yaz\tekin.udml -B D:\ALL\repository\temps\yaz\050310_BAW.rpd -O D:\ALL\repository\temps\yaz\050310_BAW.rpd

Tuesday, October 05, 2010

How to delete item by UDML

How to delete item by UDML
We are complaining about “UDML works great but we cannot delete item by using UDML”. Here is how to delete;

1.       Copy the item you want to delete as UDML

DECLARE FOLDER ATTRIBUTE "XXX - BAW Customer Service".."- Task Created By"."Created By User Name" AS "Created By User Name" UPGRADE ID 2160852698 LOGICAL ATTRIBUTE  "Core"."Dim - Created By"."User Name" OVERRIDE LOGICAL NAME
            ALIASES ("User Name")
            PRIVILEGES ( READ);

2.       And change it as below with DELETE command, save it.

DELETE FOLDER ATTRIBUTE "XXX - BAW Customer Service".."- Task Created By"."Created By User Login";

3.       Run nQUDMLexec.exe, that’s it!