Index partitioning.

Hi All,
Is it possible to partition the maxvalue index partion created using range index partitioning.

>
thats fine , it is possible for table. But is it possible for Index ?
>
Sure - the process is similar.
See the 'split_index_partition' clause of ALTER INDEX in the SQL doc
http://docs.oracle.com/cd/B12037_01/server.101/b10759/statements_1008.htm#i2136094
>
split_index_partition
Use the split_index_partition clause to split a partition of a global range-partitioned index into two partitions, adding a new partition to the index. This clause is not valid for hash-partitioned global indexes. Instead, use the add_hash_index_partition clause.
Splitting a partition marked UNUSABLE results in two partitions, both marked UNUSABLE. You must rebuild the partitions before you can use them.
Splitting a usable partition results in two partitions populated with index data. Both new partitions are usable.
>
And the doc has an example of how to do it.
>
Splitting a Partition: Example
The following statement splits partition p2 of index cost_ix (created in "Creating a Range-Partitioned Global Index: Example") into p2a and p2b:
ALTER INDEX cost_ix
SPLIT PARTITION p2 AT (1500)
INTO ( PARTITION p2a TABLESPACE tbs_01 LOGGING,
PARTITION p2b TABLESPACE tbs_02);

Similar Messages

Regarding Renaming INDEX Partitions

Hi All,
I have the following syntax for renaming table partitions/subpartitions :
ALTER TABLE SMTP_MSG_TRAFFIC_FCT RENAME PARTITION FOR (20100310) TO BASE_FACT_20100310;
ALTER TABLE SMTP_MSG_TRAFFIC_FCT RENAME SUBPARTITION FOR (20100310,19) TO BASE_FACT_20100310_MAILBOX;
I want to do the same thing for renaming the index partitions/subpartitions :
ALTER INDEX SMTP_MSG_TF_FX02_ARR_TM_DIM_ID RENAME PARTITION FOR (20100310) TO BASE_FACT_20100310;
ALTER TABLE SMTP_MSG_TF_FX02_ARR_TM_DIM_ID RENAME SUBPARTITION FOR (20100310,19) TO BASE_FACT_20100310_MAILBOX;
But I get the following error:
SQL> ALTER INDEX SMTP_MSG_TF_FX02_ARR_TM_DIM_ID RENAME PARTITION FOR (20100310) TO BASE_FACT_20100310;
ALTER INDEX SMTP_MSG_TF_FX02_ARR_TM_DIM_ID RENAME PARTITION FOR (20100310) TO BASE_FACT_20100310
ERROR at line 1:
ORA-14006: invalid partition name
Is it possible to rename the partition in this way? Or can I set any parameter that will ensure that when the base table partition name changes, it will also change the index partition name?
Thanks & Regards,
Aniket

yes its possible:
SQL>ALTER INDEX index_name RENAME PARTITION index_partiton_name TO new_index_partition;
You are getting error because :
a partition name is expected but not present as you typed,
enter an appropriate partition name, will solve the problem.
Edited by: adnanKaysar on Mar 18, 2010 12:32 PM

Data partitioning & Index partitioning

Hi,
I have the following questions about implementing ILM on a existing table consisting of millions of rows and with index partitioning...
1) how can data be partitioned when index partitioning exists?
2) what happens to the indexes when the table data is partitioned (per month)?
3) when the data is partitioned, will the indexes AUTOMATICALLY reorganize according to the data partitions? otherwise, what has to be done to the index partitioning?
Thanks.

Hi,
The best way to do this would be to create a new table with the desired partitioning and build a local index on this. If it is possible to take the table offline and recreate all its dependencies manually, then the new table can simply be renamed to the old table. If the table needs to stay online, you may want to look into using online redefinition with dbms_redefinition. Below I have shown a sample of how you could recreate the table and index:
Original definitions -
create table my_test_table (id number, val number);
create index my_test_table_idx
on my_test_table (id)
global partition by range (id)
(partition p1_ind values less than (5),
partition p2_ind values less than (maxvalue));
New definitions -
create table my_test_table_new (id number, val number)
partition by range (id)
(partition p1_tab values less than (5),
partition p2_tab values less than (maxvalue));
create index my_test_table_idx_new
on my_test_table_new (id)
local;
Renaming -
alter table my_test_table rename to my_test_table_old;
alter table my_test_table_new rename to my_test_table;
alter index my_test_table_idx rename to my_test_table_idx_old;
alter index my_test_table_idx_new rename to my_test_table_idx;

ADD PARTITION시 NEW LOCAL INDEX PARTITION에 관하여

제품 : ORACLE SERVER
작성날짜 : 2004-08-16
ADD PARTITION시 NEW LOCAL INDEX PARTITION에 관하여
==================================================
(※ Oracle Partitioning Option은 8~10g Standard Edition에서는 지원하지 않는다.)
개 요
=======
ALTER TABLE ADD PARTITION을 할 경우 table partition의 경우 table partition,
tablespace, physical storage attribute를 지정할 수 있으나,
local index partition의 경우 관련 syntax를 지원하지 않는다 (8.1.6 현재)
오라클은 자동으로 index partition, tablespace, physical storage attribute를 지정한다.
(ALTER TABLE SPLIT PARTITION도 마찬가지)
new index partition의 이름은 new table partition과 같은 이름이 할당되고,
tablespace, physical storage attribute는 local index에 지정된 default값이 사용되고,
local index에 지정된 default값이 없으면 user의 default tablespace가 사용된다.
그러므로 PARTITION 추가하여 사용하는 경우 index partition 관리에 주의가 요구된다.
다음은 ALTER TABLE ADD PARTITION한 이후에 index partition을 관리하는 예제이다.
예 제
=======
-- Create partition table --
CREATE TABLE tbl_emp (
empno NUMBER,
ename VARCHAR2(20),
job VARCHAR2(20)
PARTITION BY RANGE ( empno ) (
PARTITION pt_tbl_emp01 VALUES LESS THAN( 1000 ) TABLESPACE data_emp01,
PARTITION pt_tbl_emp02 VALUES LESS THAN( 2000 ) TABLESPACE data_emp02,
PARTITION pt_tbl_emp03 VALUES LESS THAN( 3000 ) TABLESPACE data_emp03
-- Create local index --
CREATE INDEX idx_emp ON tbl_emp( empno )
TABLESPACE users --> 여기서 local index default tablespace 지정
LOCAL (
PARTITION pt_idx_emp01 TABLESPACE indx_emp01,
PARTITION pt_idx_emp02 TABLESPACE indx_emp02,
PARTITION pt_idx_emp03 TABLESPACE indx_emp03
-- PK 설정 --
ALTER TABLE tbl_emp ADD CONSTRAINT pk_emp PRIMARY KEY ( empno );
-- Local Index를 PK index로 사용함을 확인 --
SELECT index_name FROM user_indexes WHERE table_name = 'TBL_EMP';
INDEX_NAME
IDX_EMP
DROP INDEX idx_emp
ERROR at line 1:
ORA-02429: cannot drop index used for enforcement of unique/primary key
-- Add partition --
ALTER TABLE tbl_emp
ADD PARTITION pt_tbl_emp04 VALUES LESS THAN( 4000 )
TABLESPACE data_emp04;
-- 추가된 Local index partition의 이름, Tablespace를 확인한다 --
SELECT partition_name, tablespace_name, status
FROM user_ind_partitions
ORDER BY 1;
PARTITION_NAME TABLESPACE_NAME STATUS
PT_IDX_EMP01 INDX_EMP01 USABLE
PT_IDX_EMP02 INDX_EMP02 USABLE
PT_IDX_EMP03 INDX_EMP03 USABLE
PT_TBL_EMP04 USERS USABLE
-- Tablespace는 local index의 default tablespace에 생긴다 --
SELECT index_name, def_tablespace_name
FROM user_part_indexes;
INDEX_NAME DEF_TABLESPACE_NAME
IDX_EMP USERS
--<< Local index partition name 변경 >>--
ALTER INDEX idx_emp
RENAME PARTITION pt_tbl_emp04 TO pt_idx_emp04;
--<< Local index tablespace 변경 >>--
ALTER INDEX idx_emp
REBUILD PARTITION pt_idx_emp04 TABLESPACE indx_emp04;
SELECT partition_name, tablespace_name, status
FROM user_ind_partitions
ORDER BY 1;
-- 변경된 Local index partition의 이름, Tablespace를 확인한다 --
PARTITION_NAME TABLESPACE_NAME STATUS
PT_IDX_EMP01 INDX_EMP01 USABLE
PT_IDX_EMP02 INDX_EMP02 USABLE
PT_IDX_EMP03 INDX_EMP03 USABLE
PT_IDX_EMP04 INDX_EMP04 USABLE
* 다음과 같이 local index의 default tablespace를 변경할 수도 있다.
alter index <indexname>
modify default attributes tablespace <new tablespace name>;

>
1. If i drop one partition of a table why all the local indexes(year_1998 ... year_2000) get dropped but year_idx remains in database? I couldn't find index in dba_part_indexes view but i could find only year_idx line in dba_indexes view.
>
After dropping a partition, the row in DBA_PART_INDEXES still exists, so I'm unsure why it does not appear on your system. The index however does not appear in the DBA_INDEXES view before or after the drop partition:
for example
SQL> select owner, index_name from dba_part_indexes where
2 index_name = 'YEAR_IDX';
OWNER                          INDEX_NAME
APPS                           YEAR_IDX
SQL> alter table all_facts drop partition year_1999;
Table altered.
SQL> select owner, index_name from dba_part_indexes where
2 index_name = 'YEAR_IDX';
OWNER                          INDEX_NAME
APPS                           YEAR_IDX
SQL> select owner, index_name from dba_indexes
2 where index_name = 'YEAR_INDEX';
no rows selected

Loading data to indexed partitioned table

Hi,
My table is partitioned on day wise and index is created on first column (non unique)
is it possible to disable the last partition index disable to increase the speed of data loading to last partition.
create table my_table(sr_no,sr_name,doj)
partitioned on day wise field doj (date of joining).
my table size is very high (1TB) ,

OraFighter wrote:
Is it possible to disable only last partition..?
modify_index_partition
Use the modify_index_partition clause to modify the real physical attributes, logging attribute, or storage characteristics of index partition partition or its subpartitions. For a hash-partitioned global index, the only subclause of this clause you can specify is UNUSABLE.
when all else fails Read The Fine Manual
ALTER INDEX

EQG-30281: Unable to detect index partition mode

Hello,
i tried to configure a search index run over a web source with Oracle Secure enterprise Search Server 10.1.8.4. But when the plan is executed the following error occurs:
15:58:47:712 INFO     main          Cookie support is enabled
15:58:47:712 INFO     main          Maximum number of cookies = 300
15:58:47:712 INFO     main          Maximum number of cookies per host = 20
15:58:47:712 INFO     main          Maximum size of cookie = 4096 bytes
15:58:47:712 INFO     main          Cache file deletion is disabled
15:58:47:712 INFO     main          SQL hook ID is "3"
15:58:47:712 INFO     main          SQL command hook statement is "begin ? := eq_crw.crawler_get_command(?,?,?,?); end;"
15:58:47:712 INFO     main          SQL response hook statement is "begin ? := eq_crw.crawler_send_response (?,?,?,?,?,?,?,?,?,?,?,?,?); end;"
15:58:47:712 INFO     main          Crawler launched by schedule "ocbwiki"
15:58:47:712 INFO     main          Bad title to be replaced = "powerpoint presentation"
15:58:47:712 INFO     main          Bad title to be replaced = "slide 1"
15:58:47:712 INFO     main          Zip file processing package is java.util.zip
15:58:47:712 INFO     main
15:58:47:712 INFO     main          =================== Crawling status ===================
15:58:47:712 INFO     main          Crawling started at 1/11/10 3:58 PM
15:58:54:634 INFO     main          URL manager connecting to Oracle...
15:58:54:853 INFO     main          connected
15:58:54:931 INFO     main          Queue manager connecting to Oracle...
15:58:55:150 INFO     main          connected
15:58:56:228 ERROR     main          java.sql.SQLException: ORA-06550: line 1, column 20:
PLS-00302: component 'IS_INDEX_PARTITION_ENABLED' must be declared
ORA-06550: line 1, column 7:
PL/SQL: Statement ignored
15:58:56:244 ERROR     main          EQG-30281: Unable to detect index partition mode
     at oracle.search.crawler.WebCrawler.init(WebCrawler.java:897)
     at ImtCrawler.run(ImtCrawler.java:1954)
     at ImtCrawler.main(ImtCrawler.java:534)
15:58:56:244 ERROR     main          EQG-30249: Crawler process aborted
Anyone an idea?
Regards
Stefan

I believe this error indicates a failure to properly upgrade from 10.1.8.2 to 10.1.8.4. SES has been upgraded but the underlying Oracle Text installation has not for some reason.
If this is a fresh installation, I would suggest dropping the whole thing, reinstalling 10.1.8.2 then upgrading to 10.1.8.4. If it's an installation with lots of sources already crawled, suggest you log an SR with support so they can help you figure out how to recover from this.

IQ 16 - Support index partitioning?

Dear IQ gurues,
My colleagues using Vertica and known (don't no from) that SAP (Sybase) IQ doesn't support index partitioning
Please clarify - Does SAP IQ support index partitioning or not?

Think of the FP index as the default data storage mechanism. This is no different than other databases that support partitioning but global indexes only. Not all RDBMS engines support partition based indexes.
I suggest you open a case and ask for a feature enhancement around this. It has been discussed for quite some time and I tell all customers the same thing. A case with a customer use case goes a lot further to getting features like this onto the radar screen of product management and into the product.
Mark

HANA - Support index partitioning?

Dear HANA experts,
Does SAP HANA support index partitioning or not?

There's a presentation from me available here in SCN (Further Playing with SAP HANA ) that explains indexes in the column store in more detail.

Huge file in one of the search index partition folder

Hi,
I got a very strange situation in my SP farm. I have two index partitions on two WFE servers, each has Index Partition 0 and 1. On one of them, starting yesterday a very huge file grew over 30GB in the Index Partition 0 - Journal folder. I found the same
file in other Journal folder, most of them with less than 100MB.
Anyone has insight on this kind of behavior? Thanks in advance!
Regards,
Edwin

Hi Edwin,
Please help collect the following information for further troubleshoot：
1.Check the Search Application Topology if the index partition is in normal state.
2.Check the ULS log if there are any errors when the issue happened.
Best Regards,
Wendy
TechNet Community Support
Please remember to mark the replies as answers if they help, and unmark the answers if they provide no help. If you have feedback for TechNet Support, contact
[email protected]

Best way to insert data into a non indexed partitioned table?

Hi,
I've a rather basic question that I haven't been able to find an answer to so hope that someone may be able to help.
We need to insert sales data into a partitioned table that is range partitioned on sales date. The incoming sales data can span a number of days, e.g. 31-JUL-2007 through to 04-AUG-2007 inclusive. I've come up with two approaches and would like to know which you would recommend and why?
Approach 1: Load data into each individual partition using a statement like:
INSERT /*+ APPEND */ INTO sales PARTITION (sales_20070731) SELECT * FROM sales_new WHERE sales_date = TO_DATE('31-AUG-2007','DD-MON-YYYY')
Approach 2: Load data into the entire table using a statement like:
INSERT /*+ APPEND */ INTO sales SELECT * FROM sales_new
Thanks

You need to compare both approaches by creating simple test case.
But there is no advantages in approach 1until you will have index on sales_date column.
With index these approaches are comparable and you can choose whatever best fits to your requirements.
Best Regards,
Alex

Create index partition in the table partition tablespace

Hello,
I am running a custom job that
* Creates a tablespace daily
* Creates the daily table partition in the tablespace created
* Drops the tablepartition X days old
* Drops the tablespace for that partition on X+1 day.
The above job runs perfectly, but 'm having issues with managing the indexes for these partitioned tables. In the old database (10g - Single Node), all the partitions/indexes existed in one BIG tablespace and when I imported the table creation script into the new database, I modified all the table partitions & indexes to go into their respective tablespace.
Eg:
Table_name........Partition_name.....................Index_Part_name..........................Tablespace_name
============...================............====================...........=================
TABL1...................TABL1_2012_07_16............TABL1_IDX_2012_07_16............TBS_2012_07_16
TABL1...................TABL1_2012_07_15............TABL1_IDX_2012_07_15............TBS_2012_07_15
But now when the job runs, it creates the index into the default tablespace TBS_DATA.
Table_name........Partition_name.....................Index_Part_name..........................Tablespace_name
============...================.............====================...........=================
TABL1...................TABL1_2012_08_16............TABL1_IDX_2012_08_16............TBS_DATA
TABL1...................TABL1_2012_08_15............TABL1_IDX_2012_08_15............TBS_DATA
I can issue alter index rebuild to move the indexes to its default tablespace, but how can I ensure that the index gets created in its designated tablespace?
NOTE - the partition/tablespace management job that I run only creates the table partition and not the index.
The new env is a 2-Node 11gR2 RAC cluster on Linux x86_64.
Thanks in advance,
aBBy.

Excerpt from the job -
This creates the partition into the new tablespace.
v_sql_new_part := 'alter table '||tab_owner||'.'||tab_name||' add partition '||v_new_part_nm||'
values less than (to_date('''||v_new_part_dt_formatted||''',''DD-MON-YYYY'')) tablespace '||part_tbs;
execute immediate v_sql_new_part;New tablespace for new partition - because this is a 10T database and having multiple tablespaces helps with backup/recovery.
Thanks,
aBBy.

Indexes/Partitioning on Exadata

I am converting an Oracle based Forecasting system from RAC to Eaxadata. I've been told to drop all of my indexes. Is this really wise? Also, should I un-partition my tables?

You almost certainly don't want to drop all indexes. Arup Nanda has an excellent presentation on the topic:
The Arup Nanda Blog: Collab14 Session Should You Drop Indexes in Exadata?
HTH!
Marc

OMBPlus Index Partition Scripting

Has anyone managed to script the creation of partition keys and partitions against an index that already exists?
The documentation inc scripting doc doesn't seem to help, and OMBPLUS just keeps prompting for SET or RENAME. It can be done through the UI but is very very laborious.
Any help gratefullly received.
Thanks
Mark

alter index <owner>.<index_name> parallel (degree 1); --- This is similar to noparallel clause.
Anand

Indexes, Partitions

Can someone advise on how I can see the details for indexes and partitions for a InfoCube. Thanks in advance

When you say "what the logic is behind a cubes partition number increasing rapidly", are you trying to understand the process of adding partitions for each new Request or why SAP BW does it?
As far as pitfalls, the only downside of compression (condensing, and I think I've seen some older doc refer to it as collapsing), is that once you have compressed a Request, you can not back the Request out of the cube. So you usually want to have some time period between loading a Request and when you compress it - enough time to be comfortable that the data is correct. We typically use 14 days for things that are loaded daily, but I'm sure you could find shops that compress sooner and others that wait much longer to compress.
The non-cumulative stuff,like inventory, require that marker dates be used. This often results in no reduciton in rows for these appls. There is a How To paper on how to handle them.
You'll find that when you start to have hundres of F fact table partitions for a cube, versus one or only several for the E fact table, query performance begins deteriorate.
Also - depending on your cube, you may be able to partition the E fact table by 0FISCPER or 0CALMONTH and have your queries that advantage of partition pruning to improve performance.

Objects (table, index, partition) fragmented

Hi all,
I run the script bellow to find fragmented objects:
select segment_name, segment_type, count(*) from dba_extents
where owner = ownname
group by segment_name, segment_type
having count(*) > 1
order by 3
And i get some liste like this:
object_names object_types 249
My question is the objects that i have are fragmented and what can i do to solve that fragmentation i mean to avoid object fragmentation.
regards
raistarevo

Hi,
You can go to the 8i documentation for more information:
http://download-west.oracle.com/docs/cd/A87860_01/doc/appdev.817/a76936/dbms_s2a.htm#1004668
But I advice you to read this thread below first.
http://asktom.oracle.com/pls/asktom/f?p=100:11:0::::P11_QUESTION_ID:8381899310385
Cheers

Index partitioning.

Similar Messages

Maybe you are looking for