I'm trying to find a way to ADD new partitions to local indexes and at the same time specify their tablespaces without having to DROP and RECREATE.
Here´s an example table based on yearly partitioning:
CREATE TABLE "TABL_ANOM"
(
ANOM_TS TIMESTAMP(6) NOT NULL
, ANOM_TIPO NUMBER(2, 0) NOT NULL
, ANOM_NIVEL NUMBER(2, 0) NOT NULL
, ANOM_ID NUMBER(10, 0) NOT NULL
[code]...
Here´s an index def for the table:
CREATE INDEX "TABL_ANOM_INDEX1" ON "TABL_ANOM" ("ANOM_NIVEL") LOCAL
(PARTITION DGSCOPSX_2011 TABLESPACE DGSCOPSX_2011
,PARTITION DGSCOPSX_2012 TABLESPACE DGSCOPSX_2012
)
OK. Now I want to add partitions for 2013 so for the table I use:
ALTER TABLE TABL_ANOM ADD
PARTITION DGSCOPS_2013 VALUES LESS THAN (TO_DATE('2014-01-01 00:00:00','YYYY-MM-DD HH24:MI:SS')) TABLESPACE DGSCOPS_2013;
and this works fine for the table but I can't find a similar command to simply add additional partitions to the indexes. I know that I can drop and recreate the indexes with the additional partition defs but on some of my tables, I'm dealing with hundreds of millions of rows and I think it would take way too long to drop and recreate all indexes on all partitions.
Also related is the PRIMARY KEY index partitions. Is there a way to add partitions (specifying the tablespaces) without having to DROP and re-ADD the CONSTRAINT with the additional partition for 2013?
We are evaluating partition strategies with view to achieving performance gains in reporting in particular. How efficient are partitioned indexes in this regard e.g.
just partitioned indexes on an un partitioned table.
One large fact table with durrogate keys on which have bitmpa indexes which link to unique key in associated dimensions. Considering partitioning the bitmap index which links to the largest dimension and similarly partition the dimension key on largest dimension.
And here is what I have done. I have created a procedure that takes 3 input parameters, please see the attached script:
1) TABLE_OWNER 2) TABLE_NAME 3) PARTITION_NAME - requires to query the particular partition to get the partition name
a situation where they will input dates as the partition was on a DATE column, now my challenge is how to incorporate this into the procedure to accept DATE as an input which will require one to query the particular table to get dates.I thought of using (HIGH_VALUE - 1) to get the dates from ALL_IND_PARTITIONS.
2day i was dropping few unwanted index from the data base, By mistake i removed the local partitioned index , So i want to recreate that index.i create the index, will the partitioned index updates when we add partitioned to the tables.
I have table with 4 partition by range partition. I am loading the table in bulk mode to latest partition. Before I load , I dropped the index and after Load I will be creating index. So when I am dropping index, it is dropping index from all the partitions and when creating the index, I am creating the index for all partitions. When I am creating index using local, it is telling you have to create local index for all partitions at the same time. because of that I have to drop and recreate all indexes again. Again I have to gather stats for whole table .
I was thinking we can build index for one partition and index should remain as is for old partitions If this is not the case, how do I plan my load for a partitioned table using bulk mode to latest partition.
For partitioned tables (RANGE, LIST types ) what are circumstances when a Global or a Local Index become UNUSABLE ? I was told that in some circumstances where the Indexes become UNUSABLE , not even a SELECT query against the table will work. Is this true ? For partitioned tables with Global Index, the global index will be listed in DBA_INDEXES . Right ?
writing a trigger to drop partitions with zero rows which are older than 6months and drop the local indexes and rebuild the global indexes for any schema in a databaase ?
I have tried the below code :
declare v_statement varchar2(600); v_rows number; begin for x in (select * from dba_tab_partitions [code]........
I want to avoid using row number and also want to dynamically select a schema when executing the script.
I have a huge table (about 60 gb) partition over range. The index on this table is global index created on 4 columns together. I have a query which is running very slowly. The explain plan is showing the use of this global index.Explain plan is not showing pstart and pend because the index is global.
SNRCTL> stat LISTENER Connecting to (DESCRIPTION=(ADDRESS=(PROTOCOL=IPC)(KEY=LISTENER))) STATUS of the LISTENER ------------------------ Alias LISTENER Version TNSLSNR for IBM/AIX RISC System/6000: Version 11.2.0.1.0 - Production Start Date 19-JAN-2013 00:50:10 Uptime 0 days 0 hr. 29 min. 51 sec Trace Level off Security ON: Local OS Authentication [code]....
In every oracle documentation for e.g:11.2 Scan and Node TNS Listener Setup Examples [ID 1070607.1] we found the local listener status showing both local-ip and vip. Why is not showing in our case?
I have an procedure with four in parameters in it. Now, I want to add an out parameter to it. The issue is the procedure is called from many other procedures and triggers. Since we can't set default value to out parameter as we can for in parameter, how we can achieve the requirement without having to update all the places where this procedure is being called.
I would like to add an additional column indicator (Y or N) to specify which ID's do not contain all records of 'NONE'. There can be an occurrence of 'NONE' as long as there is another cust_num different to 'NONE' These should be marked as 'Y' but in cases where all the ID's cust_num = 'NONE' only then these should be marked as 'N'.
I need some functionality in current (4.2.2) APEX websheet which seems not possible out of the box.Therefore I'd like to ad a javascript to my websheet application.
I have successfully set up a (test) environment for single-instance Grid Infrastructure and Oracle database using job role separation. So I have the recommended grid and oracle users, and the oinstall, dba, oper, asmadmin, asmdba and asmoper groups. I have the following directory structure for my Oracle Bases and Oracle Homes:
/u01/app/11.2.0/grid - GI home /u01/app/grid - GI base /u01/app/oracle - DB base /u01/app/oracle/product/11.2.0/db_1 - DB home
Platform is OEL5, 11.2.0.1 GI and DB.
This all works fine.
What I now want to do (since what I'm trying to do is make this environment as secure as practical) is set up an additional sysoper operating system account, so that that user can connect to carry out sysoper tasks, amongst other things stopping and starting the instance.
So (as the oracle user) I do the following:
$ sqlplus / as sysdba SQL> create user test identified by passwrod; SQL> grant sysoper to test; SQL> exit
As you can see, shutdown works, but startup doesn't. I have registered an SR with Oracle on this, but they're not being useful. The suggestion was that I should give testoper the to install secondary group. But this is not a secure solution as this now gives testoper privileges to do things in OB / OH that it really shouldn't be able to do. No other suggestions have been forthcoming. The reason for the suggestion however is the ownership / permissions on the Grid OH oracle executable:
[root@db03 ~]# ls -la /u01/app/11.2.0/grid/bin/oracle -rwsr-s--x 1 grid oinstall 184286237 Aug 22 11:15 /u01/app/11.2.0/grid/bin/oracle
As you can see, it has group oinstall, so you can see why giving the user oinstall group would work. But in my view this is not satisfactory.
One thing that occurred to me is that in this environment I should arguably in fact be using srvctl to stop / start instances. But that means setting up a user with asmoper role (presumably) rather than oper. That user would then (again presumable) be able to stop and start other GI resources, which is not what I want.
I would like to create a trigger on a table which populates a log table. In addition to using the table where the trigger will exist, I would like to populate a couple more fields in the log table with with data from 2 other tables.
e.g.
NAME_TABLE -reg_id -name
ADDRESS_TABLE *trigger to be fired when a new record is created here. -reg_id -srv_id
PROCESS_TABLE -srv_id -start_time -end_time
This is what I would like the logging table to look like:
I am trying to migrate a table to a new table that has the field sequence changed and also has a new field added. My main question is if it is possible to have datapump add values to the new field in the target table.For example:
-original table has fields a, b, d, c -new table has fields b, c, d, a, e
I want to load the new table and also include adding values for field e. In this case, field e is a year field, so it should be loaded with '2012'..Does datapump have the ability to do this? Is reorganizing the fields going to cause me any problems? We are on oracle version 11.2.0.3
I would like to ask about indexes in partitioned tables.I have indexes on a partitioned table, it is partitioned by range method i.e based on Creation date time.All select queries sent to the table use the Creation date time. I have an index on Creation date time.Here is an example:
SELECT col1, col2, col3 FROM table1 where date_time BETWEEN TO_DATE ('20120117 10:00:00','YYYYMMDD HH24:MI:SS') AND TO_DATE ('20120117 13:00:00','YYYYMMDD HH24:MI:SS') AND frmt_name = 'XXXX' AND sender = 'YYYYY' AND nature = 'ZZZZ' AND type LIKE '548' ORDER BY date_time WHERE ROWNUM <= 5000 [code]....
do I have to add DATE_TIME to all indexes (IX_NAME_FORMAT_TYPE,IX_CCY) or not?
CREATE TABLE CARGA_P_RECARGA_NEW1 ( TELE_NUM VARCHAR2(10) NOT NULL, FECHA DATE NOT NULL,
[Code].....
Then I tried to insert some rows in that table, every insert statement is like this:
INSERT INTO CARGA_P_RECARGA_NEW1 VALUES ('3134769595','20/01/2013 07:22:50','1107','CONFB_20121121_20121122175002 60000000000000000090.TXT',0,16,'8327--7991284',1);
Every insert I executed had the month 01 because I expected to query results only from partittion p_0113 but nevermind how query I execute, the result is always the same. I mean if I excute this statement:
SELECT * FROM CARGA_P_RECARGA_NEW1 P_0113;
I get the same result when I execute any other like this:
We recently had to delete data from the table. This was a simple delete statement with a where clause and without taking into consideration any partition/subpartition clauses. Post committing the delete we have a count mismatch problem with two queries in particular
select count(0) count_without_parallel FROM TRANSACTION_TABLE t;
--THIS RETRIEVES *15774811* ROWS
select /* parallel(t,default) */count(0) count_with_parallel FROM TRANSACTION_TABLE t+
--THIS RETRIEVES *15777617* ROWS WHICH IS THE ACTUAL EXPECTED COUNT.
I also ran the following just to summarize
select (select count_with_parallel from ( select /* parallel(t,default) */count(0) count_with_parallel FROM TRANSACTION_TABLE t))+ - +(select count_without_parallel from (+ select count(0) count_without_parallel FROM TRANSACTION_TABLE t)) as false_difference from dual;
The difference in *2806* rows as expected.To re-affirm my counts I ran
select /*+ parallel(t,default) */ 'count_on_t',count(*) from TRANSACTION_TABLE t group by 'count_on_t' order by 1;
--THIS RETRIEVES *15777617* ROWS
Removing the parallel hint reverts back to the lesser count. Not sure what is wrong but something prevents the query from parsing the whole table and/or partitions and subpartitions.
I have partitioned the table based on field.But when I am selecting by Partition or by the field I am getting Explain plan as Table Access full.I am pasting the sql and Explain Plan here. The table has two partition by BOOKING_DT_WID. One less than 20100801 and other less than 99991231.
CODESELECT * FROM WC_BOOKING_SALESREP_F WHERE BOOKING_DT_WID >= 20100801; SELECT * FROM WC_BOOKING_SALESREP_F PARTITION(SALESREP_LESS1_99991231); Here is the Explain Plan for the same. CODESELECT STATEMENT ALL_ROWSCost: 1,501 Bytes: 293,923,641 Cardinality: 809,707 4 PX COORDINATOR [code]....
How do I know if the sql is doing partition prune.
understand on the below regarding partitioning table and indices.
1. How to know if each partition have LOCAL index or Global index?
I had the above question duw to the beloe reason :
I have a table with 130 partitions and 7 indices. When checked the DBA_IND_PARTITIONS it lists down only 1 index name (which distinct) for all the 130 partitions.
Does it mean it is a Global index? If Global index, is it partitioned? How to get those information?
2. Assuming, if i have LOCAL indices for every partitions, from where i shall nail down the name of each local index and how would i get the detail of referencing Global index?
3. By which type a table was partitioned? LIST,HASH,RANGE ?
Does "Update on a Partitioned Column" cause fragmentation ?
See example below :
Suppose I have a table as below that has approx 3 million rows and growing :
1) Table ABC : file_id number status_flag varchar2(1) file_Content clob date_created date
2) This table is list partitioned on Column "status_flag".
Column "status_flag" can take 3 values "A" or "B" or "C"
3) A file_id can transition from 1 "Status_flag" to another. [ A--> B , C--> B, B-->C, A --> C, etc ]
This means its possible to UPDATE the "status_flag " - the partitioned column.
Question : ------------- 1) Would the movement of rows from 1 partition to another cause fragmentation ? 2) Or Instead, should this be achieved by maintaining 3 tables - 1 each for every "status_Flag" as Table_A, Table_B, Table_C
This would mean that, if file_id 1, changes "status_flag" from 'A' to 'B' : -- The corresponding row from Table_A will be created in Table_B --- The same will be deleted from Table_A
This would still cause fragmentation ... with the overhead of Inserting a CLOB column from 1 table to another.
3) How to determine how much percent of the table is fragmented ?
