I currently have a 5 column index on a table with over 2 billion records (paritioned on created_date (weekly) that is not very effective.I am contemplating replacing this 5 key index and creating a new single column index made up by hashing of all the 5 five columns.
Is this a wise stratgey? How can I implement this so it is most effective and I dont shoot myself in the foot?
I have in my database (OLTP-System) a table with about 6000000 records and a zise of about 2GB.
the way to create multi_column indexes on the table?
What are the rule to define the best-position of a column in an index?
index_1(col_1,Col_2,col_3) and not [ (col_1,Col_3,col_2) or (col_2,Col_3,col_1) or (col_2,Col_1,col_3) or (col_3,Col_2,col_1) or (col_3,Col_1,col_2) ] ?
how to create a primary key with out creating an index?
View 10 Replies View RelatedI have to create some indexes in a production database. Do I need to Compute Statistics after creating indexes? Or when I create they automatically are computed?
The version I'm using is:
Oracle Database 10g Release 10.2.0.5.0 - 64bit Production
PL/SQL Release 10.2.0.5.0 - Production
CORE 10.2.0.5.0 Production
TNS for 64-bit Windows: Version 10.2.0.5.0 - Production
NLSRTL Version 10.2.0.5.0 - Production
What should be the order of columns while creating b-tree non-unique index.
Low cardinality first
or
Hight cardinality first
I have a view on base tables holding historical data for previous 60 months(one table per month) with union all operators.create index on those base tables will improve performance or creating a primary key with disabled novalidate will improve for retrieving data?
The view has around 8 million data and used as a fact table with 4 dimension tables.A DTS package from MSSql side refreshes OLAP cube by retrieving data from these tables in oracle.
I've been successfully setting up text indexes on multiple columns on the same table (using MULTI_COLUMN_DATASTORE preferences), but now I have a situation with a one-to-many data collection table (with a FK to a lookup table), and I need to search columns across both of these tables. Sample code below, more of my chattering after the code block:
CREATE TABLE SUBMISSION
( SUBMISSION_ID NUMBER(10) NOT NULL,
SUBMISSION_NAME VARCHAR2(100) NOT NULL
);
CREATE TABLE ADVISOR_TYPE
[code]...
I've looked at DETAIL_DATASTORE and USER_DATASTORE, but the examples in Oracle Docs for DETAIL_DATASTORE leave me a little bit perplexed. It seems like this should be pretty straightforward. I'm trying to avoid creating new columns, and keeping the trigger adjustments to a minimum.
During developing Domain Index for Oracle 11.2.0.1.0 (problem also appears in 12с) i was faced with misunderstanding of parameter types of function ODCIIndexInsert in case of creating indextype WITH ARRAY DML option.According to Oracle documentation [URL}...
In case of WITH ARRAY DML option Oracle will invoke ODCIIndexInsert with following signature
FUNCTION ODCIIndexInsert(
ia ODCIIndexInfo,
ridlist ODCIRidList,
newvallist varray_of_column_type,
env ODCIEnv)
RETURN NUMBER
In my case indexed column has datatype NUMBER so i defined varray_of_column_type as SYS.ODCINumberList STATIC FUNCTION ODCIIndexInsert(ia in sys.ODCIIndexInfo, ridlist in sys.ODCIRidList, newvallist in sys.ODCINumberList, env in SYS.ODCIEnv) RETURN NUMBER
Indextype was created as
CREATE INDEXTYPE test_index_type
FOR
test_eq(number, number)
USING index_methods
WITH ARRAY DML(number, sys.ODCINumberList)
[code]...
(problem occurs in all cases)
CREATE TABLE test_table (id NUMBER (19,0));
CREATE INDEX test_index ON test_table(id) INDEXTYPE IS test_index_type;
When attempting to insert data in the table insert into test_table values (1);
oracle raise exception
CODEError starting at line 53 in command:
insert into test_table values (1)
Error at Command Line:53 Column:1
Error report:
[code]...
So my question is.Is it normal behavior of oracle (according to documentation)?What is correct signature of ODCIIndexInsert function in case of INDEXTYPE creation with 'WITH ARRAY DML' option and fact that indexed column has NUMBER datatype?
By the way if i define indextype without 'WITH ARRAY DML' option signature is clear, and working. But this approach doesn't satisfies our performance needs.Also if i define index type with option 'WITH ARRAY DML WITHOUT COLUMN DATA' and use signature static function ODCIIndexInsert(ia sys.odciindexinfo, ridlist sys.odciridlist, env sys.ODCIEnv) return number
Everything works too. But this approach doesn't satisfies our business needs.Is it a way to define ODCIIndexInsert parameter types (in case of indexing number column) so that batch inserting works according to documentation ?
FUNCTION ODCIIndexInsert(
ia ODCIIndexInfo,
ridlist ODCIRidList,
newvallist varray_of_column_type,
env ODCIEnv)
I am attaching full sql script to recreate environment and reproduce the problem.Type definition:
CREATE OR REPLACE TYPE index_methods AS OBJECT
(
step number,
STATIC FUNCTION ODCIGetInterfaces(ifclist OUT SYS.ODCIObjectList) RETURN NUMBER,
STATIC FUNCTION ODCIIndexCreate (ia SYS.ODCIIndexInfo, parms VARCHAR2, env SYS.ODCIEnv) RETURN NUMBER,
STATIC FUNCTION ODCIIndexAlter (ia sys.ODCIIndexInfo, parms IN OUT VARCHAR2, altopt number, env sys.ODCIEnv)
[code]...
is there some performance/access difference between a bitmap index on a number column and char(1) column? Both columns are not null with a default value.My application has a querie like this:
select ass.column20, ass.column30from table_a pucinner join table_b asson ass.column1 = puc.column1where pc.column_char = 'S'and ass.column_char02 = 'P'
If I create a bitmap index on column "column_char", the access plan is not changed. But changing the column datatype to number(1) and obviously the values, the index is accessed and the cost decreases.This table has 4.000.000 rows. Oracle 11.2.0.2SO
While creating the index we are getting the error "ORA-00603: ORACLE server session terminated by fatal error".
We have the space in tablespaces and also in the file systesm.
I have a table with column A which contains very few null values. I need to select these rows. I am considering two options:
a) create function based index on NVL(A, 0) and use this in where clause NVL(A, 0)=0 (column doesn't have values 0)
b) create function based index on NVL2(A, 0, NULL) and and use this in where clause NVL2(A, 0, NULL) = 0
First idea was option A. But I realized in option B the index will be much smaller, because most of values of column A isn't NULL so NVL2 will return NULL and index will not have as much leafs as in NVL. It is good idea to use NVL2? Is there any against to use option B instead of A?
I have created one unique index on one column of my table. Now i would like to add one more column in the same index without dropping the index.
SQL > CREATE TABLE DEBUG_TABLE
2 (
3 SLNO NUMBER,
4 MESSAGE VARCHAR2(4000 BYTE),
5 CREATED_DATE DATE DEFAULT SYSDATE,
6 CREATED_TIME TIMESTAMP(6) DEFAULT SYSDATE
7 );
Table created.
SQL > CREATE UNIQUE INDEX index_debug1 ON debug_table (SLNO);
Index created.
SQL > ALTER INDEX index_debug1 ADD COLUMN MESSAGE;
ALTER INDEX index_debug1 ADD COLUMN MESSAGE
*
ERROR at line 1:
ORA-02243: invalid ALTER INDEX or ALTER MATERIALIZED VIEW option
SQL >
If my query is under execution and I want to make an index on a column which is very much needed by my query. Will a simple index solve the purpose or is there any extra keyword required ?
View 2 Replies View RelatedIs the index suggested on the date column.
Here is the nature of the date column in my case:
1. The table populates with 1000's of records every day with date being always incremental (current date).
2. The search criteria from the weba application (ADF) is based on the date, user gives the range.
3. From ADF I am referring to it as sql.Timestamp when building the query.
Does Index suggested on the date column here and if so what type of index ?
I have a table say MY_TAB with columns as below
emp_id number,
name varchar2(30),
from_dt date,
remarks varchar2(60)
insert into MY_TAB values (1,'TOM','01-JAN-13', 'some remark');
insert into MY_TAB values (1,'TOM','02-JAN-13', 'some remark');
insert into MY_TAB values (2,'TOM','01-JAN-13', 'some remark');
insert into MY_TAB values (3,'TOM','01-JAN-13', 'some remark');
insert into MY_TAB values (4,'TOM','01-JAN-13', 'some remark');
insert into MY_TAB values (4,'TOM','02-JAN-13', 'some remark');
How do I ensure that when a user tries to insert record with emp_id as 1, then he should only be allowed to enter another from_dt but the value in the name column have to be the same as in the previous row of emp_id 1.
insert into MY_TAB values (1,'TOOM','03-JAN-13') --shld not be allowed.
I'm getting an error as follows .
create table asgnd_agent_bak as (Select * from ASGND_AGENT a, SCN s
where
a.CNTCT_KEY = s.CNTCT_KEY and a.SCN_NUM=s.SCN_NUM
and a.ACTVTY_DT = to_date('03/17/2013','mm/dd/yyyy')
and s.SCN_OPEN_DT = to_date('03/15/2013','mm/dd/yyyy')
and a.SRC_SYS_DESC = 'FACET');
create table asgnd_agent_bak as (Select * from ASGND_AGENT a, SCN s
*
ERROR at line 1:
ORA-00957: duplicate column name
I have a table with, for example, three columns: A, B,C.
I execute on this table only one select: CODESELECT * FROM TABLE WHERE A = :1 AND B=:2
Column A has a lot of distinct values (numbers), but B can have only two values: 'Y' or 'N' (cardinality about 50%/50%). It is worth to create index on two columns: A, B? Does query using index on A column will be much slower than using index on A, B?
I am creating an index in program and then drop the index at the end of the program. Some times due to some problem if the index could not be dropped and the user again runs the program then we get the error
ORA-01408 Index already exist on the column.
how I can get away with this error or how I can check whether the index with the same columns exits prior to creating an index.
Is there any command like
Create or replace index.....
what my issue is with this virtual column
CREATE TABLE C0HARPA.VCOL_TAB
(
col1 VARCHAR2(30 BYTE),
v_col1 VARCHAR2(6) GENERATED ALWAYS AS (SUBSTR(col1,1,6)) VIRTUAL
)
TABLESPACE TOOLS
[code]...
Lastly, can an index be created on a virtual column? I am thinking no since it calculated?
Version 4.0.2.00.07
I have an Interactive Report with a column that I need to create a link on but I need to pass the value of another column.
This report shows the Features of a particular Release. There is a select list for the Release that is selected and executes the report. There is a column that indicates whether the Feature has a Commitment attached to it. If there is have that a link to another report that shows all the Commitments for that Feature by passing the FEATURE_ID which is also a column on the Feature report.
Query:
SELECT
....
,CASE WHEN ft.feature_id IN (SELECT cc.feature_id
FROM customer_commitments cc
WHERE ft.feature_id = cc.feature_id
) THEN 'Yes' ELSE 'No'
END "Commitment Exists"
,ft.feature_id
....
FROM ....If "Commitment Exists" has a 'Yes', the 'Yes' is a link that will pass the FEATURE_ID to another report that will show the Commitments for that Feature. If the "Commitment Exists" has a 'No' then no link. I was also wondering if the Commitment report could open in a seperate window and make it modal?
explain slow performance of multicolumn indexes on oracle 11g R2 observed in the following scenario? A multi-column index (b-tree index) not partitioned, not unique, not reversed with 3 columns.
A series of queries are run using all 3 columns. The performance hit comes when the first order column values changes. So, maybe after 10 select queries the value changes. The 2nd and 3rd order columns are changing throughout the series of select calls, but no performance bottleneck it hit then.
here is my query
SELECT CURRENTSTEP
FROM (SELECT ( WFENTRY.NAME
|| ','
|| CURRENTSTEP.STEP_ID
) AS CURRENTSTEP,
(CASE
WHEN WFENTRY.NAME IN
[Code]...
in this query I am concatenating tow columns , I use this query as a sub query in my other queries and filter the results with and CURRENTSTEP = ?
here is how I use it
select
sys_audit_id
from
( SELECT
*
FROM
(SELECT
F.FINDING_NUMBER,
[Code]....
I saw adding this as a subquery with the filter and CURRENTSTEP = ? is slowing my query very much , as this is a derived column i cannot add index then how to improve performance for this subquery ?
I have a table A with a column B timestamp(6). The tables contains around 300000 rows..I have created index 'idx' on the column B.When i compare column 'B' with systimestamp, it does not use the index, whereas if i compare 'B' with sysdate it uses the index.
Eg :
select count(*) from a where b<=sysdate;
The above used the index 'idx' and executed in 1 second
select count(*) from a where b<=systimestamp;
The above does not use the index and executed in 19 seconds.
Version Info: 11.2 on Solaris 10
I have a partitioned table like below. I want to create a B-Tree index on SALES_RGN column which is neither the part of Primary key or the Partitioned key. Should I create this index as local or Global ?
CREATE TABLE sales_dtl
(
txn_id number (9),
salesman_id number(5),
salesman_name varchar2(30),
sales_rgn varchar2(10), -----------------------------> This column needs to be indexed
sales_amount number(10),
sales_date date,
constraint pk_sales_dtl primary key (txn_id)
[code]....
I have a report with 2 groups Gheader and Glines.The report looks at PO headers and lines. I want to create a data link from the the 2 queries based on the line id in po_lines_all.However I only want to select this in the lines query so I do not get repeating records at the header query.
View 11 Replies View RelatedWe have a table called address and having the address fields and city ,state etc. The table will store huge amount of data .We need to query on the table. I would like to know how can we fasten the query and improve the performance of the query by creating index on these columns...Query is given below . note that the nullable columns can have data
SELECT *
FROM address
WHERE address1 = 'a'
[Code]....
I'm altering a column length to increase the size and getting "ORA-30556: functional index is defined on the column to be modified".
On searching more about this error, it seems like the function index must be dropped before altering the column.The table I'm dealing with is huge.
Question 1:In case of dropping and recreating the index, should the following steps be done:
- Drop Index
- Alter the column to increase the size
- Recreate the index with NOLOGGING and NOPARALLEL clause
- ALTER INDEX to turn on LOGGING
- Gather Statistics on that index
Question 2:Is there anything else that should be done when the index is dropped and re-created?
Question 3:What are the side-effects of carrying out the above steps in a huge table with around 15 million rows?
Question 4:Would it work if I disable the index, alter the column and reenable the index?Do I have to rebuild the index and gather Stats upon reenabling it?
I need to create a composite unique index on varchar2, number and CLOB column. I haven't used such index before that have the CLOB column indexing. I found the below link related to CLOB indexing...
[URL]......
Links from where I can get related info. Also I would like to know the impact of such index on performance. I have to store and process around 50 million records in such a way, will it be beneficial to use this index?
I have an index on column of table which of data type varchar2. While selecting data from that table I am using following scenarios in where on the indexed column
like '%abc%'
like 'abc%'
like '&abc'
Will be the corresponding index will be for those cases?
I am working with following select clause:
select distinct S.ID ID
from
ods.hso_Scheduled H,
ods.SO_SCHEDULED S
where
S.insertion_date >= to_date('01-DEC-2011') and S.insertion_date < to_date('01-FEB-2012')
and H.ID=S.ID
Both the involved tables, HSO_SCHEDULED is having 15 million records and SO_SCHEDULED table is having 7 million records.
I have created following indexes on these tables:
Indexes on SO_SCHEDULED:
Index name Column name
SS_IDX1ID, SO_SUB_ITEM__ID
SS_IDX2INSERTION_DATE
SS_IDX3ID, INSERTION_DATE
SS_IDX4ID, SO_SUB_ITEM__ID, INSERTION_DATE
SO_SCHEDULED_ID_PKID
Indexes on HSO_SCHEDULED:
HSS_IDX1ID, SO_SUB_ITEM__ID, LAST_UPDATING_DATE
HSS_IDX2ID, LAST_UPDATING_DATE
HSS_IDX3ID
My problem is despite of having relevant indexes present, my query is not hitting them and hence the performance is very bad.
Explain Plan:
Execution Plan
----------------------------------------------------------
Plan hash value: 574170360
--------------------------------------------------------------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes |TempSpc| Cost (%CPU)| Time | TQ |IN-OUT| PQ Distrib |
--------------------------------------------------------------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 814K| 38M| | 9574 (1)| 00:02:15 | | | |
| 1 | PX COORDINATOR | | | | | | | | | |
| 2 | PX SEND QC (RANDOM) | :TQ10002 | 814K| 38M| | 9574 (1)| 00:02:15 | Q1,02 | P->S | QC (RAND) |
| 3 | HASH UNIQUE | | 814K| 38M| 185M| 9574 (1)| 00:02:15 | Q1,02 | PCWP | |
|* 4 | HASH JOIN | | 2653K| 124M| | 9564 (1)| 00:02:14 | Q1,02 | PCWP | |
| 5 | PX JOIN FILTER CREATE| :BF0000 | 814K| 22M| | 3903 (1)| 00:00:55 | Q1,02 | PCWP | |
| 6 | PX RECEIVE | | 814K| 22M| | 3903 (1)| 00:00:55 | Q1,02 | PCWP | |
| 7 | PX SEND HASH | :TQ10000 | 814K| 22M| | 3903 (1)| 00:00:55 | Q1,00 | P->P | HASH |
| 8 | PX BLOCK ITERATOR | | 814K| 22M| | 3903 (1)| 00:00:55 | Q1,00 | PCWC | |
|* 9 | TABLE ACCESS FULL| SO_SCHEDULED | 814K| 22M| | 3903 (1)| 00:00:55 | Q1,00 | PCWP | |
| 10 | PX RECEIVE | | 14M| 272M| | 5654 (1)| 00:01:20 | Q1,02 | PCWP | |
| 11 | PX SEND HASH | :TQ10001 | 14M| 272M| | 5654 (1)| 00:01:20 | Q1,01 | P->P | HASH |
| 12 | PX JOIN FILTER USE | :BF0000 | 14M| 272M| | 5654 (1)| 00:01:20 | Q1,01 | PCWP | |
| 13 | PX BLOCK ITERATOR | | 14M| 272M| | 5654 (1)| 00:01:20 | Q1,01 | PCWC | |
| 14 | TABLE ACCESS FULL| HSO_SCHEDULED | 14M| 272M| | 5654 (1)| 00:01:20 | Q1,01 | PCWP | |
--------------------------------------------------------------------------------------------------------------------------------
Predicate Information (identified by operation id):
---------------------------------------------------
4 - access("H"."ID"="S"."ID")
9 - filter("S"."INSERTION_DATE">=TO_DATE(' 2011-12-01 00:00:00', 'syyyy-mm-dd hh24:mi:ss') AND
"S"."INSERTION_DATE"<TO_DATE(' 2012-02-01 00:00:00', 'syyyy-mm-dd hh24:mi:ss'))