Limit on number of function based indexes ?

What is the limit on the number of function based indexes that can exist on a table ?
I haven't been able to find this info in the Oracle docs yet . I'd RTFM, but which one ? There's so many ! :-)

http://download-east.oracle.com/docs/cd/B19306_01/server.102/b14237/limits003.htm#sthref4186

Similar Messages

Function based indexes on CLOB storage

On a 10gR2 database, with schema-less CLOB storage for an XMLType column:
(1) Can a function based index include a wildcard in the namespace ? Or do I need a new function based index for each specific namespace ?
(2) I must create a new function based index for each different element that I want an indexed search on ?
(3) What limit is there on the number of function based indexes per table ?
(4) I believe XQuery can include a wildcard for namespaces, but XPath 1.x can't. Can I create a function based index using XQuery, rather than XPath ?
Documents conforming to different versions of an XML schema will be present (schema versioning), but I want to search across all documents irrespective of a specific namespace - e.g. "Find any document with reference = 'some Value' , and amount = 1000".
CLOB storage is proposed, due to the need to handle documents from multiple versions of an XML schema. The knowledge of the XSD is not known at development time, but is user definable, and it must be possible to change the structure without system down time. Structured storage is not suitable, due to Oracle's requirement for downtime if the schema changes (CopyEvolve drops/recreates tables), and Oracle doesn't support schema collections, so you can't bind an XML column to multiple schemas.
Here is some sample code of what I'm trying to do:
create table BulkTest
ID NUMBER(10) not null primary key,
USERFIELDS XMLTYPE
create sequence S_BulkTest;
--Document conforming to version 1 of schema
INSERT INTO BulkTest(id, Userfields) VALUES
(S_BulkTest.Nextval,
'<?xml version="1.0" encoding="utf-8" ?>
<mt395 xmlns="urn:incident:mt395-1">
     <reference>FH12345678</reference>
     <relatedReference>FH23456789</relatedReference>
     <queries>Here is some query text.</queries>
     <narrative>Here is some narrative text.</narrative>
     <relatedMessageType>300</relatedMessageType>
     <relatedMessageDate>2005-03-29</relatedMessageDate>
     <direction>R</direction>
     <sessionNumber>1234</sessionNumber>
     <isn>123456</isn>
     <relatedMessageDescription>This is the deal where I bought USD 1 million for GBP at 1.76.</relatedMessageDescription>
     <otherParty>232332</otherParty>
</mt395>'
--Document conforming to version 2 of schema
INSERT INTO BulkTest(id, Userfields) VALUES
(S_BulkTest.Nextval,
'<?xml version="1.0" encoding="utf-8" ?>
<mt395 xmlns="urn:incident:mt395-2">
     <guid>0f9a08f6-b052-4693-baba-8f7dc881e7e8</guid>
     <reference>333333</reference>
     <queries>Another query</queries>
     <narrative>Some narrative</narrative>
     <direction>R</direction>
     <sessionNumber>1234</sessionNumber>
     <isn>223456</isn>
</mt395>'
--It seems I need to create a new index for each field I want to search on
create index iBulkTest_REFERENCE
on BulkTest
(extractValue(UserFields,'/mt395/reference', 'xmlns="urn:incident:mt395-1"'));
--And that a new index is required for each specifc namespace that is present
--Can't we include a wildcard in the namespace ?
create index iBulkTest_REFERENCE_2
on BulkTest
(extractValue(UserFields,'/mt395/reference', 'xmlns="urn:incident:mt395-2"'));
--If I want to query, I have to explicitly specify each namespace.
--Can't I specify a wildcard ?
--This will make it "fun" querying across namespaces!
select
id,
extractValue(UserFields,'/mt395/reference', 'xmlns="urn:incident:mt395-2"') As Reference,
t.userfields.getclobval() userfields
from bulktest t
WHERE extractValue(UserFields,'/mt395/reference', 'xmlns="urn:incident:mt395-2"') = '333333'

Andy
#1. You do not have scehma versioning here. Your model is totally incorrect. You shoud not change the namespace when versioning the XML Schema. You have 2 different and totally disjoint XML Schemas. The correct was to version, as distinct from evolve an XML Schema is to change the Schema Location Hint associated with your XML...
Eg
INSERT INTO BulkTest(id, Userfields) VALUES
(S_BulkTest.Nextval,
'<?xml version="1.0" encoding="utf-8" ?>
<mt395 xmlns="urn:incident:mt395" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="urn:incident:mt395 mt395-1.xsd">
<reference>FH12345678</reference>
<relatedReference>FH23456789</relatedReference>
<queries>Here is some query text.</queries>
<narrative>Here is some narrative text.</narrative>
<relatedMessageType>300</relatedMessageType>
<relatedMessageDate>2005-03-29</relatedMessageDate>
<direction>R</direction>
<sessionNumber>1234</sessionNumber>
<isn>123456</isn>
<relatedMessageDescription>This is the deal where I bought USD 1 million for GBP at 1.76.</relatedMessageDescription>
<otherParty>232332</otherParty>
</mt395>'
--Document conforming to version 2 of schema
INSERT INTO BulkTest(id, Userfields) VALUES
(S_BulkTest.Nextval,
'<?xml version="1.0" encoding="utf-8" ?>
<mt395 xmlns="urn:incident:mt395 xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="urn:incident:mt395 mt395-2.xsd">
<guid>0f9a08f6-b052-4693-baba-8f7dc881e7e8</guid>
<reference>333333</reference>
<queries>Another query</queries>
<narrative>Some narrative</narrative>
<direction>R</direction>
<sessionNumber>1234</sessionNumber>
<isn>223456</isn>
</mt395>'
This is the correct way of versioning an XML Schema. THe namespace stays the same, the SchemaLocationHint in the SchemaLocation tag changes.
Bear in mind that if you use the technique you are currently using you will make any path expressions you need to write absolutely unmaintaining and the processing of them very inefficient.
Some questions to consider
Node 'X' in namespace 'X' is never the same as Node 'X' in namespace 'Y'.
How would you write an Xpath or XQuery that targetted multiple versions, but not all versions ?
What happens if you have other documents that are really in a different namespace ? Using wildcards can you differentiate them..
From the problem you are describing and the terminolgy you are using it looks like you've been an early customer of Yukon. MSFT clearly didn't understand schema versioning in the early beta releases and used the 'change the namespace' schema for modelling schema versioning.
We do have some technology coming down the pipe which can address the issue, regardless of whether or not it is too late for you to correct the versioning scheme you have selected. However I cannot discuss that in a public forum. If you want to learn about these features and are prepared to enter an NDA with Oracle in order to do so please contact me directly. You can do this a number of ways...
Guess my email address @oracle.com
Post your email address here and I'll delete the post as soon as I have it..
Update your OTN Forum profile to include your email address
Open a TAR and post the tar number here. You can then softclose the tar as this is simply a method for me to get your contact info.

Function-based index with OR in the wher-clause

We have some problems with functin-based indexes and
the or-condition in a where-clause.
--We use Oracle 8i (8.1.7)
create table TPERSON(ID number(10),NAME varchar2(20),...);
create index I_NORMAL_TPERSON_NAME on TPERSON(NAME);
create index I_FUNCTION_TPERSON_NAME on TPERSON(UPPER(NAME));
The following two statements run very fast on a large table
and the execution-plan asure the usage of the indexes
(-while the session is appropriate configured and the table is analyzed):
1)     select count(ID) FROM TPERSON where upper(NAME) like 'MIL%';
2)     select count(ID) from TPERSON where NAME like 'Mil%' or (3=5);
In particular we see that a normal index is used while the where-clause contains
an OR-CONDITION.
But if we try the similarly select-statement
3)     select count(ID) FROM TPERSON where upper(NAME) like 'MIL%' or (3=5);
the CBO will not use the function-index I_FUNCTION_TPERSON_NAME and we have a full table scan in the execution-plan.
(This behavior we only expect with views but not with indexes.)
We ask for an advice like a hint, which enable the CBO-usage
of function-based indexes in connection with OR.
This problem seems to be artificial because it contains this dummy logic:
     or (3=5).
This steams from an prepared statement, where this kind of boolean
flag reduce the amount of different select-statements needed for
covering the hole business-logic, while using bind-variables for the
concrete query-parameters.
A more realistic (still boild down) version of our select-statement is:
select * FROM TPERSON
where (upper(NAME) like 'MIL%' or (NAME is null))
and (upper(FIRSTNAME) like 'MICH% or (FIRSTNAME is null))
and ...;
thank you for time..
email: [email protected]

In the realistic statement you write :
select * FROM TPERSON
where (upper(NAME) like 'MIL%' or (NAME is null))
and (upper(FIRSTNAME) like 'MICH% or (FIRSTNAME is null))
and ...;
as far as i know, NULL values are not indexed, "or (NAME is NULL)" have to generate a full table scan.
HTH
We have some problems with functin-based indexes and
the or-condition in a where-clause.
--We use Oracle 8i (8.1.7)
create table TPERSON(ID number(10),NAME varchar2(20),...);
create index I_NORMAL_TPERSON_NAME on TPERSON(NAME);
create index I_FUNCTION_TPERSON_NAME on TPERSON(UPPER(NAME));
The following two statements run very fast on a large table
and the execution-plan asure the usage of the indexes
(-while the session is appropriate configured and the table is analyzed):
1)     select count(ID) FROM TPERSON where upper(NAME) like 'MIL%';
2)     select count(ID) from TPERSON where NAME like 'Mil%' or (3=5);
In particular we see that a normal index is used while the where-clause contains
an OR-CONDITION.
But if we try the similarly select-statement
3)     select count(ID) FROM TPERSON where upper(NAME) like 'MIL%' or (3=5);
the CBO will not use the function-index I_FUNCTION_TPERSON_NAME and we have a full table scan in the execution-plan.
(This behavior we only expect with views but not with indexes.)
We ask for an advice like a hint, which enable the CBO-usage
of function-based indexes in connection with OR.
This problem seems to be artificial because it contains this dummy logic:
     or (3=5).
This steams from an prepared statement, where this kind of boolean
flag reduce the amount of different select-statements needed for
covering the hole business-logic, while using bind-variables for the
concrete query-parameters.
A more realistic (still boild down) version of our select-statement is:
select * FROM TPERSON
where (upper(NAME) like 'MIL%' or (NAME is null))
and (upper(FIRSTNAME) like 'MICH% or (FIRSTNAME is null))
and ...;
thank you for time..
email: [email protected]

Function-based Index and an OR-condition in the WHERE-clause

We have some problems with functin-based indexes and
the or-condition in a where-clause.
(We use oracle 8i (8.1.7))
create table TPERSON(ID number(10),NAME varchar2(20),...);
create index I_NORMAL_TPERSON_NAME on TPERSON(NAME);
create index I_FUNCTION_TPERSON_NAME on TPERSON(UPPER(NAME));
The following two statements run very fast on a large table
and the execution-plan asure the usage of the indexes
(-while the session is appropriate configured and the table is analyzed):
1)     select count(ID) FROM TPERSON where upper(NAME) like 'MIL%';
2)     select count(ID) from TPERSON where NAME like 'Mil%' or (3=5);
In particular we see that a normal index is used while the where-clause contains
an OR-CONDITION.
But if we try the similarly select-statement
3)     select count(ID) FROM TPERSON where upper(NAME) like 'MIL%' or (3=5);
the CBO will not use the function-index.
(This behavior we only expect with views but not with indexes.)
We ask for an advice like an hint, which enable the CBO-usage
of function-based indexes in connection with OR.
This problem seems to be artificial because it contains this dummy logic:
     or (3=5).
This steams from an prepared statement, where this kind of boolean
flag reduce the amount of different select-statements needed for
covering the hole business-logic, while using bind-variables for the
concrete query-parameters.
A more realistic (still boild down) version of our prepared select-statement run in
SQL Plus:
define x_name = 'MIL%';
define x_firstname = '';
select * FROM TPERSON
where (upper(NAME) like '&x_name' or ( '&x_name' = ''))
and (upper(FIRSTNAME) like '&x_firstname' or ('&x_firstname' = ''))
and ...;
In particular we dont refernce the tablecolumn , but the QUERY-Parameter
yield the second boolean value in the or-condition.
The problem is that this condition ('&x_name' = '') dont use any index.
thanks a lot for spending your time with this problem

Try
SELECT /*+ RULE */
as your hint. I don't have the book with me, but this last weekend I read a section about your very problem. The book was a Oracle Press gold cover about Oracle 8i Performance tuning. If you e-mail me I can quote you the chapter when I get home Friday.

Function Based Index on Date Column

Hi All,
I need to execute a query like this :
SELECT * FROM ORDERS WHERE APPROVE_DATE IS NULL
I read anywhere that this will cause unnecessary FTS so that I should create function based index.
I have tried one below , but not sure that this is correct approach :
CREATE INDEX idx_1
ON ORDERS (NVL(APPROVE_DATE, '01-JAN-1900'));
SELECT * FROM ORDERS WHERE NVL(APPROVE_DATE, '01-JAN-1900') = '01-JAN-1900'
Is this a correct approach ?
Thank you,
xtanto

A SQL_TRACE output will explain clearly what Justin has stated.
I have created a table T based on all_objects.
SQL> desc t
Name                                      Null?    Type
OWNER                                     NOT NULL VARCHAR2(30)
OBJECT_NAME                               NOT NULL VARCHAR2(30)
SUBOBJECT_NAME                                     VARCHAR2(30)
OBJECT_ID                                 NOT NULL NUMBER
DATA_OBJECT_ID                                     NUMBER
OBJECT_TYPE                                        VARCHAR2(19)
CREATED                                            DATE
LAST_DDL_TIME                             NOT NULL DATE
TIMESTAMP                                          VARCHAR2(19)
STATUS                                             VARCHAR2(7)
TEMPORARY                                          VARCHAR2(1)
GENERATED                                          VARCHAR2(1)
SECONDARY                                          VARCHAR2(1)
CASE I_
SQL> select count(1) from t
2 /
COUNT(1)
    934320
SQL> select count(1) from t where created is null
2 /
COUNT(1)
      2376The number of null values in CREATED column is proportionately very small.
Now i execute the query without function based index.
select *
from t
where created is null
call     count       cpu    elapsed       disk      query    current        rows
Parse        1      0.00       0.09          0          0          0           0
Execute      1      0.00       0.00          0          0          0           0
Fetch      160      0.04       0.10          0      12662          0        2376
total      162      0.04       0.19          0      12662          0        2376
Rows     Execution Plan
      0 SELECT STATEMENT   GOAL: ALL_ROWS
   2376   TABLE ACCESS   GOAL: ANALYZED (FULL) OF 'T' (TABLE)And here is the query that uses the function based index
select *
from t
where nvl(created,to_date('01-01-1900','DD-MM-YYYY')) = to_date('01-01-1900','DD-MM-YYYY')
call     count       cpu    elapsed       disk      query    current        rows
Parse        1      0.01       0.00          0          0          0           0
Execute      1      0.00       0.00          0          0          0           0
Fetch      160      0.01       0.01          0        698          0        2376
total      162      0.03       0.01          0        698          0        2376
Rows     Execution Plan
      0 SELECT STATEMENT   GOAL: ALL_ROWS
   2376   TABLE ACCESS   GOAL: ANALYZED (BY INDEX ROWID) OF 'T' (TABLE)
   2376    INDEX   GOAL: ANALYZED (RANGE SCAN) OF 'T_FN_IDX' (INDEX)Its very obvious from the above output that the Function Based Index as increased the performance.
CASE II_
SQL> select count(1) from t
2 /
COUNT(1)
    934320
SQL> select count(1) from t where created is null
2 /
COUNT(1)
    202168Now the null values in the CREATED column is proportionately large than the first test case.
Now lets see without using the function based index
select *
from t
where created is null
call     count       cpu    elapsed       disk      query    current        rows
Parse        1      0.00       0.00          0          0          0           0
Execute      1      0.00       0.00          0          0          0           0
Fetch    13479      0.46       0.71          2      25832          0      202168
total    13481      0.46       0.71          2      25832          0      202168
Rows     Execution Plan
      0 SELECT STATEMENT   GOAL: ALL_ROWS
202168   TABLE ACCESS   GOAL: ANALYZED (FULL) OF 'T' (TABLE)Now iam trying to use the function based index
select *
from t
where nvl(created,to_date('01-01-1900','DD-MM-YYYY')) = to_date('01-01-1900','DD-MM-YYYY')
call     count       cpu    elapsed       disk      query    current        rows
Parse        1      0.00       0.00          0          0          0           0
Execute      1      0.00       0.00          0          0          0           0
Fetch    13479      0.54       0.84          0      33826          0      202168
total    13481      0.54       0.84          0      33826          0      202168
Rows     Execution Plan
      0 SELECT STATEMENT   GOAL: ALL_ROWS
202168   TABLE ACCESS   GOAL: ANALYZED (FULL) OF 'T' (TABLE)Its obvious from the result that oracle has decided to go for a FULL TABLE SCAN even when an index was available.
So just having a function based index is not going to increase the query performance. There are lot of other factors to be considered as stated above.
Thanks,
Karthick.

Function Based Index MOD(column_name,8)

For a function based index (MOD(column_name,8), the query is not using it. Do anyone have a solution to it? The table is having millions of rows.

Oracle only uses indexes if the number of returned rows is less than (i think) 4%. In your case the number of rows is 12,5% so Oracle will never use this index unless you force it

Function-based indexes don't seem to work in Oracle 8.1.5?

Hi,
What gives? What am I doing wrong? I have a table AIRPORT with a column (varchar2(64)) which I have specified a function based index for, but I can't get SQL wueries to use it!!!! the following SQL executes a FULL TABLE SCAN:
select /*+ index (a idx_upper_cityname) */ *
from airport a
where nls_upper(cityName) = 'dfdf'
...as does...
select *
from airport a
where nls_upper(cityName) = 'dfdf'
Table and index code is as follows:
CREATE TABLE airport
id NUMBER NOT NULL,
citycode VARCHAR2(3) NOT NULL,
cityname VARCHAR2(64) NOT NULL,
state VARCHAR2(2),
country VARCHAR2(2) NOT NULL,
region CHAR(1),
airportcode VARCHAR2(3) NOT NULL,
airportname VARCHAR2(64),
code VARCHAR2(4)
drop index idx_upper_cityname
CREATE INDEX idx_upper_cityname ON airport nls_upper(substr(cityName, 0, 64) )
Environment is as follows:
Oracle8i v8.1.5 running on WinNT v4.0 (SP 5)
Client is running on the same machine
thanks in advance,
Alexander

New data point: when I set the handler in my logging.properties file thusly,
org.apache.catalina.core.ContainerBase.[Catalina].[info-dev].[/infoisland].level = ALL
org.apache.catalina.core.ContainerBase.[Catalina].[info-dev].[/infoisland].handlers = java.util.logging.ConsoleHandlerI get 0 bytes in the info-dev log (which used to have the aforementioned expception in it). Where is my console going?

Problem using two function based indexes at once!

Hello Oracle!
I've got problems using two function based indexes on geometries at once.
The problem occures, when I use a spatial join between two geometries both using function based indexes.
The test case:
CREATE TABLE quad (centroid NUMBER);
CREATE TABLE points (no NUMBER, point MDSYS.SDO_GEOMETRY);
CREATE OR REPLACE FUNCTION getQuad (centroid NUMBER) RETURN MDSYS.SDO_GEOMETRY DETERMINISTIC IS
BEGIN
RETURN MDSYS.SDO_GEOMETRY(2003, NULL, NULL, MDSYS.SDO_ELEM_INFO_ARRAY(1,1003,1),MDSYS.SDO_ORDINATE_ARRAY(centroid-5,centroid-5,centroid+5,centroid-5,centroid+5,centroid+5,centroid-5,centroid+5,centroid-5,centroid-5));
END;
INSERT INTO USER_SDO_GEOM_METADATA VALUES('quad','tiedge.getQuad(centroid)',MDSYS.SDO_DIM_ARRAY(MDSYS.SDO_DIM_ELEMENT('X', -100, 100, .0000001), MDSYS.SDO_DIM_ELEMENT('Y', -100, 100, .0000001)),NULL);
CREATE INDEX quad_idx on quad(getQuad(centroid)) INDEXTYPE IS MDSYS.SPATIAL_INDEX;
INSERT INTO quad VALUES (0);
INSERT INTO quad VALUES (5);
INSERT INTO quad VALUES (10);
INSERT INTO points VALUES (1, MDSYS.SDO_GEOMETRY(1001,NULL,NULL,MDSYS.SDO_ELEM_INFO_ARRAY(1,1,1),MDSYS.SDO_ORDINATE_ARRAY(4,4)));
ALTER SESSION SET QUERY_REWRITE_INTEGRITY=TRUSTED;
ALTER SESSION SET QUERY_REWRITE_ENA[i]Long postings are being truncated to ~1 kB at this time.

hi there,
For a better audience for this question, I'd look at the database forum.
guys on that will be a lot more familiar with FBIs
thanks
Barry

ORA-04091 (table string.string is mutating) and Function-Based Index

I've encountered a problem with DELETEing from a table when that table has a function-based index on it. The following demonstrates this:
SQL> CREATE OR REPLACE FUNCTION get_employee_location(p_empno IN number)
2 RETURN varchar2
3 DETERMINISTIC
4 IS
5 l_return_value   varchar2(20);
6 BEGIN
7 SELECT loc
8 INTO   l_return_value
9 FROM   dept
10 WHERE deptno = (SELECT
11                   e.deptno
12                   FROM emp e
13                   WHERE empno = p_empno);
14 return l_return_value;
15 end;
16 /
Function created.
SQL> create index location_idx on emp (get_employee_location(empno));
Index created.
SQL> delete from emp;
delete from emp
ERROR at line 1:
ORA-04091: table SCOTT.EMP is mutating, trigger/function may not see it
ORA-06512: at "SCOTT.GET_EMPLOYEE_LOCATION", line 7------------------------------------------------
The question is: How can I successfully DELETE FROM emp but keep my function-based index in place?
Thanks
Andy

'Being able to' is 'being able to', but
it is dangerous to declare "DETERMINISTIC" for non-deterministic function.
The following problem happens on non-deterministic function index.
SQL> update dept set loc = 'NEWYORK' where deptno=10;
1 row updated.
SQL> commit;
Commit complete.
SQL> select * from emp where get_employee_location(deptno)='NEWYORK';
no rows selected
SQL> select * from dept;
    DEPTNO DNAME          LOC
        10 ACCOUNTING     NEWYORK
        20 RESEARCH       DALLAS
        30 SALES          CHICAGO
        40 OPERATIONS     BOSTON
SQL> select empno,ename from emp where get_employee_location(deptno)='NEW YORK';
     EMPNO ENAME
      7782 CLARK
      7839 KING
      7934 MILLER
SQL> select empno,ename,get_employee_location(deptno) from emp where deptno=10;
     EMPNO ENAME      GET_EMPLOYEE_LOCATION(DEPTNO)
      7782 CLARK
      7839 KING
      7934 MILLER
SQL> select empno,ename,get_employee_location(deptno) from emp where get_employee_location(deptno)='NEW YORK';
     EMPNO ENAME      GET_EMPLOYEE_LOCATION(DEPTNO)
      7782 CLARK      NEW YORK
      7839 KING       NEW YORK
      7934 MILLER     NEW YORK
SQL> drop index location_idx ;
Index dropped.
SQL> select empno,ename from emp where get_employee_location(deptno)='NEW YORK';
no rows selected

Problem creating a function based index

Hi,
Database: 8.0.6
OS: win 2000 server
I want to create a function based index as shown below:
sql> create index new_index on A_INP_PATIENTS_ADMT(bus_unit, admit_status_flag, nvl(clinical_discharge_yn,'N'), ward_code)
When i press return its showing error,
"ORA-00907: missing right parenthesis"
Please show me how to rewrite this query to build function based index.
Best Regards,
Edited by: ateeqrahman on Jun 6, 2010 2:34 PM

Your sql is fine:
SQL> CREATE TABLE A_INP_PATIENTS_ADMT(
2                                   bus_unit number,
3                                   admit_status_flag varchar2(1),
4                                   clinical_discharge_yn varchar2(1),
5                                   ward_code varchar2(10)
6                                  )
7 /
Table created.
SQL> create index new_index on A_INP_PATIENTS_ADMT(bus_unit, admit_status_flag, nvl(clinical_discharge_yn,'N'), ward_code)
2 /
Index created.
SQL> You did not post Oracle version. Most likely, you are on some older version that does not support FBI. Otherwise, post your version and table description and create index statement execution showing all errors you are getting.
SY.

Function Based Index And Selectivity

Hi All,
I have some doubts w.r.t FBI.I am on 10gR2 (10.2.0.4) with Solaris 5.9
I was under impression that FBI does not provide guaranteed index access and CBO choose access pattern purely on basis of available stats and query selectivity.
However, many a times i found that CBO is going for FBI even in case when FTS provides better query elapsed time.
I created following test case:
create table fbi_test (id number,flag varchar2(1));
begin
for i in 1..1000000
loop
insert into fbi_test values(i,'Y');
end loop;
end;
commit;
begin
for i in 1..10
loop
insert into fbi_test values(i,'N');
end loop;
end;
commit;
ANALYZE TABLE FBI_TEST COMPUTE STATISTICS;
CREATE INDEX fbi_test_FBI
ON fbi_test (CASE WHEN flag = 'Y' THEN 1 ELSE NULL END);
Autotrace for FBI ACCESS
SQL> select *from fbi_test where (CASE WHEN flag = 'Y' THEN 1 ELSE NULL END)=1;
1000000 rows selected.
Elapsed: 00:00:18.43
Execution Plan
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)|
| 0 | SELECT STATEMENT | | 10000 | 50000 | 342 (1)|
| 1 | TABLE ACCESS BY INDEX ROWID| FBI_TEST | 10000 | 50000 | 342 (1)|
|* 2 | INDEX RANGE SCAN | FBI_TEST_FBI | 4000 | | 1958 (1)|
Statistics
0 recursive calls
0 db block gets
136812 consistent gets
0 physical reads
0 redo size
22180292 bytes sent via SQL*Net to client
733814 bytes received via SQL*Net from client
66668 SQL*Net roundtrips to/from client
0 sorts (memory)
0 sorts (disk)
1000000 rows processed
Autotrace for FTS
SQL> select *from fbi_test where flag = 'Y';
1000000 rows selected.
Elapsed: 00:00:16.56
Execution Plan
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)|
| 0 | SELECT STATEMENT | | 500K| 2441K| 371 (9)|
|* 1 | TABLE ACCESS FULL| FBI_TEST | 500K| 2441K| 371 (9)|
Statistics
0 recursive calls
0 db block gets
68372 consistent gets
0 physical reads
0 redo size
22180292 bytes sent via SQL*Net to client
733814 bytes received via SQL*Net from client
66668 SQL*Net roundtrips to/from client
0 sorts (memory)
0 sorts (disk)
1000000 rows processed
FYI...
SQL> show parameter opt
NAME TYPE VALUE
filesystemio_options string asynch
object_cache_optimal_size integer 102400
optimizer_dynamic_sampling integer 2
optimizer_features_enable string 10.2.0.3
optimizer_index_caching integer 0
optimizer_index_cost_adj integer 100
optimizer_mode string ALL_ROWS
optimizer_secure_view_merging boolean TRUE
plsql_optimize_level integer 2
My questions are,
1.why oracle optimizer is going for FBI with high cardinality of 100000 rows?
2.Why cost of FTS is high (371) as compare to FBI (342) eventhough FTS is having fewer IO (68372) + Less Elapsed Time?
3.Why Optimizer is considering ELAPSED TIME during plan generation?
Any inpute would be highly appreciated.

user635930 wrote:
Hi All,
I have some doubts w.r.t FBI.I am on 10gR2 (10.2.0.4) with Solaris 5.9
Execution Plan
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)|
| 0 | SELECT STATEMENT | | 10000 | 50000 | 342 (1)|
| 1 | TABLE ACCESS BY INDEX ROWID| FBI_TEST | 10000 | 50000 | 342 (1)|
|* 2 | INDEX RANGE SCAN | FBI_TEST_FBI | 4000 | | 1958 (1)|
---------------------------------------------------------------------------------You're seeing three different effects here.
First - for the table access by index, Oracle has "lost" the cost of the index range scan - notice that the total cost of the query is 342, but the cost of the index access is 1958. The total cost of the query should be 2,300 and Oracle should have chosen the full tablescan automatically.
Second, the stats on the index show just one distinct value for "distinct_keys", and the optimizer has decided (for no reason I can think of - it may be a bug) to assume a 0.4% selectivity on the index.
Third, the estimated cardinality of the table and index lines differs. Whatever Oracle has done in the index line has been forgotten, and the cardinality of the table line has been based on the predicate given by your case statement and, as a "complex function", that predicate has been given a selectivity of 1% - hence the 10,000 rows estimate.
The combination of unsuitable statistics, an extreme case, and a couple of quirks in the optimizer mean that the chosen path is clearly unsuitable.
[Addendum]: It just occurred to me that part of the problem is that you collected stats on the table before you created the index. Given you're running 10g, the 'create index' would automatically generate index stats at the same time - but since it's a function-based index, there's a "virtual column" created for the table as well, and that column won't have any statistics on it - which is why you get the "fixed percentage" selectivities.
Regards
Jonathan Lewis
http://jonathanlewis.wordpress.com
http://www.jlcomp.demon.co.uk
"Science is more than a body of knowledge; it is a way of thinking" Carl Sagan
Edited by: Jonathan Lewis on Nov 29, 2008 1:15 PM

Error creating function based index

i keep getting ORA-13203 errors when creating a function-based index. The function is owned by another schema but the the user creating the function has execute privileges
[email protected]> create index mpp_nc_sidx on MPP_NONCOMPLETE(GIS.DD832UTM(LON*-1,LAT))
2 indextype is mdsys.spatial_index
3 parameters('tablespace=sde4_idx sdo_indx_dims=2');
create index mpp_nc_sidx on MPP_NONCOMPLETE(GIS.DD832UTM(LON*-1,LAT))
ERROR at line 1:
ORA-29855: error occurred in the execution of ODCIINDEXCREATE routine
ORA-13203: failed to read USER_SDO_GEOM_METADATA table
ORA-13203: failed to read USER_SDO_GEOM_METADATA table
ORA-06512: at "MDSYS.SDO_INDEX_METHOD_9I", line 7
ORA-06512: at line 1
[email protected]> select table_name,column_name from user_sdo_geom_metadata;
TABLE_NAME COLUMN_NAME
MPP_NONCOMPLETE GIS.DD832UTM(LON*-1,LAT)
[email protected]> select GIS.DD832UTM(LON*-1,LAT) from mpp_noncomplete where rownum < 3;
GIS.DD832UTM(LON*-1,LAT)(SDO_GTYPE, SDO_SRID, SDO_POINT(X, Y, Z), SDO_ELEM_INFO, SDO_ORDINATES)
SDO_GEOMETRY(2001, 82212, SDO_POINT_TYPE(864941.804, 3916953.95, NULL), NULL, NULL)
SDO_GEOMETRY(2001, 82212, SDO_POINT_TYPE(568560.541, 4181497.56, NULL), NULL, NULL)
[email protected]> select text from all_source where name = 'DD832UTM';
TEXT
FUNCTION dd832utm(x number, y number)
RETURN mdsys.sdo_geometry DETERMINISTIC
IS
geom mdsys.sdo_geometry;
BEGIN
geom := sdo_cs.transform
(mdsys.sdo_geometry (2001,8265,mdsys.sdo_point_type
(x,y, null),null,null),82212);
return geom;
END;
any help appreciated
--kassim

Hi, try to use a view:
create or replace view v_dd832utm as
select
mdsys.sdo_geometry(2001,8265,
mdsys.sdo_point_type((LON*-1),LAT, null),null,null),
82212)as GEOMETRY
from
mpp_noncomplete;
provide metadata for that view (column: GEOMETRY) and create a spatial index. your way is more sophisticated ;o)
regards, Andreas

Performance problem on function-based index

Hi guys,
I am having performance problems with the addition of new function-based indexes.
alter session set nls_comp='ANSI';
alter session set nls_sort='BINARY_CI';
* have to run this because the of case-insensitivity requirements
I have a view. for ex:
create or replace view view1
as
select * from emp1,user
where emp1.empno=user.empno
union
select * from emp2,user
where emp2.empno=user.empno
union
select * from emp3,user
where emp3.empno=user.empno and so on
When I run this it works with a full table scan. Then when i created a function-based index:
create index user_ix on
user(nlssort(empno,'NLS_SORT=BINARY_CI'));
analyze index user_ix compute statistics;
analyze table user compute statistics;
the view hangs. but when i run the individual select statements it works.
Do you guys have any idea on what's going on? Any advise is greatly appreciated.
Thanks.

LC is absolutely right. Brain cramp on my part.
On the other hand, I can't seem to coerce Oracle to apply a to_binary_double conversion as part of an implicit conversion.
var bin_dbl binary_double;
select to_binary_double(14) into :bin_dbl from dual;
SCOTT @ nx102 JCAVE9420> select * from emp where empno = :bin_dbl;
no rows selected
Elapsed: 00:00:00.14
Execution Plan
Plan hash value: 2949544139
| Id | Operation                   | Name   | Rows | Bytes | Cost (%CPU)| Time     |
|   0 | SELECT STATEMENT            |        |     1 |    39 |     1   (0)| 00:00:01 |
|   1 | TABLE ACCESS BY INDEX ROWID| EMP    |     1 |    39 |     1   (0)| 00:00:01 |
|* 2 |   INDEX UNIQUE SCAN         | PK_EMP |     1 |       |     0   (0)| 00:00:01 |
Predicate Information (identified by operation id):
   2 - access("EMPNO"=TO_NUMBER(:BIN_DBL))I'd expect that Oracle would try to convert the binary double to a number, not the other way around.
Justin

Function Based Index - Query Performance

HI,
Good Day to All..
I'd like to use function based indexes on following column(to_char(ps.user_pc_id)).
Whereas this column is part of PRIMARY KEY.
Is it possible to create a function based index on PRIMARY KEY Column?
Attached below is the query with the explain plan ...
TO_CHAR Expression - Performance
Thanks for your reply.

DTYLER_APP@pssdev2> create table dt_fbi_pk(id varchar2(20));
Table created.
DTYLER_APP@pssdev2> drop table dt_fbi_pk;
Table dropped.
DTYLER_APP@pssdev2> create table dt_fbi(id number);
Table created.
DTYLER_APP@pssdev2> create index dt_fbi_idx on dt_fbi(to_char(id));
Index created.
DTYLER_APP@pssdev2> alter table dt_fbi add constraint dt_fbi_pk primary key (id);
Table altered.
DTYLER_APP@pssdev2> select constraint_name,constraint_type, index_name from user_constraints where table_name='DT_FBI';
CONSTRAINT_NAME                C INDEX_NAME
DT_FBI_PK                      P DT_FBI_PK
1 row selected.When we created the primary key constraint, Oracle created a new index rather than using the existing one because....
DTYLER_APP@pssdev2> alter table dt_fbi drop primary key;
Table altered.
DTYLER_APP@pssdev2> select index_name from user_indexes where table_name ='DT_FBI';
INDEX_NAME
DT_FBI_IDX
1 row selected.
DTYLER_APP@pssdev2> alter table dt_fbi add constraint dt_fbi_pk primary key (id) using index dt_fbi_idx;
alter table dt_fbi add constraint dt_fbi_pk primary key (id) using index dt_fbi_idx
ERROR at line 1:
ORA-14196: Specified index cannot be used to enforce the constraint.
DTYLER_APP@pssdev2> alter table dt_fbi add constraint dt_fbi_uk unique(id) using index dt_fbi_idx;
alter table dt_fbi add constraint dt_fbi_uk unique(id) using index dt_fbi_idx
ERROR at line 1:
ORA-14196: Specified index cannot be used to enforce the constraint.We can't use a function based index to enforce a unique or primary key constraint. Changing the syntax does not help..
DTYLER_APP@pssdev2> alter table dt_fbi add constraint dt_fbi_uk unique(TO_CHAR(id)) using index dt_fbi_idx;
alter table dt_fbi add constraint dt_fbi_uk unique(TO_CHAR(id)) using index dt_fbi_idx
ERROR at line 1:
ORA-00904: : invalid identifierWe can create a unique index however
DTYLER_APP@pssdev2> drop index dt_fbi_idx;
Index dropped.
DTYLER_APP@pssdev2> create unique index dt_fbi_idx on dt_fbi(to_char(id));
Index created.but we still can't use it to enforce a unique or primary key constraint
DTYLER_APP@pssdev2> alter table dt_fbi add constraint dt_fbi_pk primary key (id) using index dt_fbi_idx;
alter table dt_fbi add constraint dt_fbi_pk primary key (id) using index dt_fbi_idx
ERROR at line 1:
ORA-14196: Specified index cannot be used to enforce the constraint.
DTYLER_APP@pssdev2> alter table dt_fbi add constraint dt_fbi_uk unique(id) using index dt_fbi_idx;
alter table dt_fbi add constraint dt_fbi_uk unique(id) using index dt_fbi_idx
ERROR at line 1:
ORA-14196: Specified index cannot be used to enforce the constraint.So no, you can't use it for a primary key. If you just want to enforce uniqueness then yes, you can do it with a unique index, but not a constraint.
DTYLER_APP@pssdev2> select * from v$version;
BANNER
Oracle Database 10g Enterprise Edition Release 10.2.0.4.0 - 64bi
PL/SQL Release 10.2.0.4.0 - Production
CORE    10.2.0.4.0      Production
TNS for Linux: Version 10.2.0.4.0 - Production
NLSRTL Version 10.2.0.4.0 - Production
5 rows selected.HTH
Daviid

How to create function based index on REGEXP_LIKE funtion

Dear Gurus,
I have below table CDR
Name Null Type
STARTTIME NOT NULL DATE
SUBSCRIBERNUMBER NOT NULL NUMBER
CALLINGNUMBER NOT NULL VARCHAR2(20)
CALLEDNUMBER NOT NULL VARCHAR2(20)
I am regularly firing below query
SELECT count(*)
FROM CDR data
WHERE STARTTIME BETWEEN '01-Jul-2009 00:00:00' and '31-May-2012 23:59:59'
AND REGEXP_LIKE(data.SUBSCRIBERNUMBER, '^98721[0-9]*[5]+[0-9]*$');
since there is REGEXP_LIKE is being used, Can I use function based index to improve performance.
Thanking in advance
Sanjeev

Hi,
you can do it that way :Scott@my11g SQL>create table test (name varchar2(30));
Table created.
Scott@my11g SQL>create index myfbi on test(case when regexp_like(name,'^98721[0-9]*[5]+[0-9]*$') then 1 else 0 end);
Index created.
Scott@my11g SQL>explain plan for
2 select * from test where case when regexp_like(name,'^98721[0-9]*[5]+[0-9]*$') then 1 else 0 end = 1;
Explained.
Scott@my11g SQL>/
PLAN_TABLE_OUTPUT
Plan hash value: 140237472
| Id | Operation                   | Name | Rows | Bytes | Cost (%CPU)| Time     |
|   0 | SELECT STATEMENT            |       |     1 |    20 |     1   (0)| 00:00:01 |
|   1 | TABLE ACCESS BY INDEX ROWID| TEST |     1 |    20 |     1   (0)| 00:00:01 |
|* 2 |   INDEX RANGE SCAN          | MYFBI |     1 |       |     1   (0)| 00:00:01 |
Predicate Information (identified by operation id):
   2 - access(CASE WHEN REGEXP_LIKE ("NAME",'^98721[0-9]*[5]+[0-9]*$')
              THEN 1 ELSE 0 END =1)
Note
   - dynamic sampling used for this statement (level=2)
19 rows selected.

Limit on number of function based indexes ?

Similar Messages

Maybe you are looking for