What is 'RANK RANK PLACE' in sql plan?

Hi all.

The oracle is 11.2.0.3 on a linux machine.

I did ' select (index ffs) parallel and insert serial.

Here is the among the tracefiles tkprof output (parallel slave produced several trace file).
This is why block related counters are not accurate.

I found 'RANK RANK PLACE' operation and I Googled it.
There is little information on this subject.

The index 'PARTITIONED_AAA_PK' is hash global index is partitioned (level 16 hash).

What is "RANK RANK PLACE"?

insert into TEST_AAA_HIS
  select /*+ index_ffs(t PARTITIONED_AAA_HIS_PK) parallel_index(t,PARTITIONED_AAA_HIS_PK,8) */ *
   from PARTITIONED_AAA t
 where seq > 548215475

call     count       cpu    elapsed       disk      query    current        rows
------- ------  -------- ---------- ---------- ---------- ----------  ----------
Parse        1      0.01       0.08         33         33          0           0
Execute      1     23.63      32.67      44336     112722     399430     1407251
Fetch        0      0.00       0.00          0          0          0           0
------- ------  -------- ---------- ---------- ---------- ----------  ----------
total        2     23.64      32.75      44369     112755     399430     1407251

Misses in library cache during parse: 1
Optimizer mode: ALL_ROWS
Parsing user id: SYS
Number of plan statistics captured: 1

Rows (1st) Rows (avg) Rows (max)  Row Source Operation
---------- ---------- ----------  ---------------------------------------------------
         0          0          0  LOAD TABLE CONVENTIONAL  (cr=112722 pr=44336 pw=0 time=32670190 us)
   1407251    1407251    1407251   PX COORDINATOR  (cr=48 pr=0 pw=0 time=4328843 us)
         0          0          0    PX SEND QC (RANDOM) :TQ10000 (cr=0 pr=0 pw=0 time=0 us cost=151381 size=93444240 card=210460)
         0          0          0     PX PARTITION HASH ALL PARTITION: 1 16 (cr=0 pr=0 pw=0 time=0 us cost=151381 size=93444240 card=210460)
         0          0          0      TABLE ACCESS BY GLOBAL INDEX ROWID PARTITIONED_AAA PARTITION: ROW LOCATION ROW LOCATION (cr=0 pr=0 pw=0 time=0 us cost=151381 size=93444240 card=210460)
         0          0          0       INDEX RANGE SCAN PARTITIONED_AAA_PK PARTITION: 1 16 (cr=0 pr=0 pw=0 time=0 us cost=731 size=0 card=210460)(object id 814571)
Thanks in advance.
Best regards.

869578 wrote:

I did ' select (index ffs) parallel and insert serial.

Here is the among the tracefiles tkprof output (parallel slave produced several trace file).
This is why block related counters are not accurate.

I found 'RANK RANK PLACE' operation and I Googled it.
There is little information on this subject.

The index 'PARTITIONED_AAA_PK' is hash global index is partitioned (level 16 hash).

What is "RANK RANK PLACE"?

For partitioned queries the plan still shows a 'start of partition' as 'partition In the case of a scan of systematic index range or comprehensive analysis of a partitioned global index, the operation at the table of visits by rowid saves these as 'PLACE of LINE'.

select /*+ index_ffs(t PARTITIONED_AAA_HIS_PK) parallel_index(t,PARTITIONED_AAA_HIS_PK,8) */ *

0          0          0       INDEX RANGE SCAN PARTITIONED_AAA_PK PARTITION: 1 16 (cr=0 pr=0 pw=0 time=0 us cost=731 size=0 card=210460)(object id 814571)

Your index_ffs() index failed - possibly because you have the wrong name in the indication of the KP index.

Concerning
Jonathan Lewis

Tags: Database

Similar Questions

  • What is the new method of SQL joins

    Dear all,

    What is the new method of SQL joins

    1

    SELECT <fields>
      FROM TableA a INNER JOIN TableB b ON a.key = b.key
    

    2

    SELECT <fields>
      FROM TableA a, TableB b
     WHERE a.key = b.key
    

    Above the two queries work very well in PL/SQL, but request that no 1 does not work in Oracle Forms 10 g

    Kind regards

    Zafar Iqbal

    Hello

    Zafar Iqbal wrote:

    Dear all,

    What is the new method of SQL joins

    1

    1. SELECT
    2. TableA, TableB INNER JOIN a and b WE a.key, b.key =

    2

    1. SELECT
    2. FROM TableA, TableB b
    3. WHERE a.key, b.key =

    Above the two queries work very well in PL/SQL, but request that no 1 does not work in Oracle Forms 10 g

    Kind regards

    Zafar Iqbal

    What, exactly, is your question?

    Literally ask you which of these 2 methods is new?  As mentioned above, none is really new.  Both participated in Oracle at least 13 years.

    Query 1, above, ANSI join syntax uses, which is more recent.   As mentioned above, Oracle has started to support that the syntax in version 9.1 (2001) previous versions only supported syntax used in query 2.

    Did you mean to ask something else?  Maybe "How can I use a query as a query 1 forms?" or "are there other places where a method works, but the other is not?

  • Setting sql plans

    Hello, I'm on 11 GR 2 and I was wondering if there is way I could plan attached to a sql_id sql based on a hash value of the plan? What I have is a sql with a sql_id which has two hash values of plan attached to the same sql_id and plan with hash value a1 is ok plan annexed to the sql_id even with the hash value a2 is unoptimal, with high cost and sql is wrong with this hash values. So I will try to find, is there a way to correct a good plan for a sql_id based on a hash value?

    To do this, you can use a SQL Plan basis or a SQL profile.

    For a profile of SQL, see script COE_XFR_SQL_PROFILE. SQL (Oracle Support id doc 215187.1) SQLT that takes a sql id and the hash value of plan that you want to apply.

  • SQL PLAN Management

    In Oracle 11 g, I want to use sql plan management to achieve stability of the regime. My problem is that I don't know what all the plans, users can run, then I put

    OPTIMIZER_CAPTURE_SQL_PLAN_BASELINES=TRUE

    But then sysaux tablespace keeps growing. I started with 2GB and now I have SYSAUX tablespace of 12GB. Is there anyway to say that only plans in certain parsing schema are automatically captured and no recursive statements to data dictionary etc. are captured. I am only interested in sql plans run by my program and not those sql statements internally generated by Oracle. I want all of this to be automatic, that’s why I set above parameter to TRUE. If that is not the right approach, what other options I have.

    The simple answer is probably that if you want to capture base lines of sql plan for a subset of users or programs, and then set the parameter to the level of the system to false and use a logon trigger to true for sessions with the subset of attributes you have defined.

    However, depending on the exact version of 11 g, there are a number of bugs affecting both the performance and the storage of the basic mechanism worth overlap Oracel version with lists of related bugs in the support Web site.

    Also double check on the occupants of the tablespace SYSAUX because the base lines are not the only thing in there that can use a significant amount of space.

  • Issue of SQL Plan Management

    Hello

    I have a question concerning the presence of 'Accepted' indicator (values or NOT) that is displayed by the function dbms_xplan.display_sql_plan_baseline.

    It is my understanding of the documentation of Oracle SQL plan does not become part of a plan SQL database until it is accepted by the process of evolution or manual loading in the base line. How is it possible to have a plan to be part of a reference that is not accepted? If the dbms_xplan.display_sql_plan_baseline function is used to display more than plans that are in a reference database and that are only in the story?


    Thanks for your thoughts on this.

    Richard

    PS. I use Oracle 11.1.0.7

    Welcome to the forum!
    >
    How is it possible to have a plan to be part of a reference that is not accepted? If the dbms_xplan.display_sql_plan_baseline function is used to display more than plans that are in a reference database and that are only in the story?
    >
    View code that you use to call the function and the result you get.

    Depending on how you call the function, and what plans are available, you will get different results. For example
    >
    plan_name
    Name of plan. It identifies a specific plan. Value default NULL means that all the plans associated with identified SQL statement are explained and displayed
    >
    You specify this setting? If this isn't the case, then all THE plans could include plans that have not been accepted.

    Section 15.5 display SQL Plan baseline in the Performance Tuning Guide contains an example of this and displays two plans: an accepted and the other not.
    http://docs.Oracle.com/CD/E11882_01/server.112/e16638/optplanmgmt.htm

    It's really the history display that includes all plans.

  • How to age a sql plan?

    Hello
    I have a query that has a sql plan. Is there a way to age on the denunciations of the analysis of the cache or cause oracle to recalculate the plan sql?

    Insaponata wrote:
    I can not make structural changes... there at - it another way?

    I don't understand you not entirely, but the runtime of CBO, check if there may be a more effective plan if yes then it will generate the new plan. Even rinse you all the plans of the shared pool (using ALTER SYSTEM EMPTY POOL SHARED) then oracle will make analysis difficult when running query and then generated new and old plan will be even (if it is not happen over the case). What exactly is your want to do? Please explain more.

  • SQL plan management unpack error

    Hi all

    Facing a problem in SQL Plan Management.

    Packed dba_sql_plan_baselines table in table called "test_baseline_14july" and exported from the source schema.

    Subsequently imported the same table in the target, now when the following error unpacking based m

    SQL > DECLARE
    number of my_plans;
    BEGIN
    my_plans: = DBMS_SPM. UNPACK_STGTAB_BASELINE (table_name = > 'test_baseline_14july');
    END;
    /

    ORA-06502: PL/SQL: digital error or value
    ORA-06512: at "SYS." DBMS_SMB', line 1024
    ORA-06512: at "SYS." DBMS_SPM', line 2933
    ORA-06512: at line 5


    The source DB: 11.2.0.2
    Destination DB: 11.1.0.6

    Let me know what is the problem?

    Thanks in advance!

    All it seems a little weird. You can think of to raise a ticket with the Support of Oracle.

    In the source database, can you show the output of

    SELECT * FROM v$version;
    SELECT * FROM dba_sql_plan_baselines;
    SELECT * FROM test_baseline_15july_2;
    

    In the destination database, can you show the output of

    SELECT * FROM v$version;
    SELECT * FROM test_baseline_15july_2;
    
  • WHAT APPLICATIONS CAN BE DEVELOPED FROM SQL SERVER 2012

    I need to know what applications can be developed using SQL SERVER 2012.

    Hi, Chris,.

    This is an overview of SQL Server 2012

    http://www.Microsoft.com/sqlserver/en/us/product-info/overview-capabilities.aspx

    SQL Server 2012 application development

    http://social.technet.Microsoft.com/wiki/contents/articles/6982.SQL-Server-2012-Developer-Training-Kit-BOM-en-us.aspx#Module_5_SQL_Server_2012_Application_Development

    Post your query on TechNet Forum to get help

    http://TechNet.Microsoft.com/en-us/SQLServer/ff898410.aspx

  • What is the date of end SQL server 2008 SP2 support?

    What is the date of end SQL server 2008 SP2 support?

    Have you tried Google? or:

    ASK THE QUESTION IN THE SQL SERVER FORUM:
    http://social.msdn.Microsoft.com/forums/en/category/SQLServer

  • What is the best place to call the HomeScreen.setNewState method?

    Hello

    I have a requirement where I need to remove the 'Asterix' around my candidacy. So what is the best place to keep the code in HomeScreen.setNewState (false) which has been determined by an electronic mail message?

    THX

    Srinivas

    You want to delete when you know that your applicaitn has been invoked, so I consider this "reset" in the activate method (() of UiAppplication)

  • What is the best place in windows 7 to discover this netwerk is slow

    Hello

    What is the best place to find out if the netwerk is slow, which is sysmonitor?

    I find myself little complicated netwerk monitor utility.

    Thank you

    Johan

    You are welcome.

    Network Monitor 3.4 is a Protocol Analyzer. It allows you to capture network traffic and view and analyze.

    How to use Network Monitor to capture network traffic

    http://support.Microsoft.com/kb/812953/en-us

    http://TechNet.Microsoft.com/en-us/library/cc938655.aspx

    Working with Network Monitor part 1

    http://www.windowsnetworking.com/articles-tutorials/trouble/working-with-network-monitor-part1.html

    Resource monitor is used to display information about the use of hardware (CPU, memory, disk and network) and software (handles files and modules) resources in real time.

    Resource monitor strongly relies on event for Windows (ETW) tracking facilities, the installation of the counter program (event tracking session) used by the resource monitor can be used for the registration of institutions

    Using the resource monitor to troubleshoot Windows Performance part 1

    http://blogs.technet.com/b/askperf/archive/2012/02/01/using-resource-monitor-to-troubleshoot-Windows-performance-issues-part-1.aspx

    Part 2

    http://blogs.technet.com/b/askperf/archive/2012/02/08/using-resource-monitor-to-troubleshoot-Windows-performance-issues-part-2.aspx

  • What is the latest version of SQL Developer?

    Hi all

    What is the latest version of SQL Developer?

    Also let me know is free or licensed?

    It can be downloaded for free from the link below.

    Oracle SQL Developer downloads

  • What metrics to use for memory capacity planning? Active or consumed (in request or use vROPS)

    Hey all,.

    I was wondering once more what metrics to use for memory capacity planning. Since the information from different sources (for example your ESXi host needs more RAM to Iwan Rahabok or VMworld 2012: Session VSP1729 - virtualized memory understanding performance management - Eric Sloof - NTPRO.NL Kit Colbert) pointing all assets (vROPS request), I expect that when we have high memory consumption / use in vCenter there will be no balloon / compression / permutation.

    Also the metric active / request should not be used for all workloads (e.g. Java / DB), why is - this? I have not found the reason of this anywhere, only the fact.

    Other sources such as Mark Achtemichuk suggests using consumed/use, which is the opposite (Understanding Active Memory - VMware vSphere Blog - Articles VMware vSphere).

    So I thought I'll with active / demand as metric and ignoring the warnings inside of vCenter.

    But the last time we have fixed half of our group, we had consumed/high memory usage and the vCenter began to inflate, compress, and finally to Exchange. Not a lot but I had not expected that, because memory active was only a small percentage of consumed.

    Can someone explain to me why this happened and who is right on this issue.

    Kind regards

    jengl

    jengl I don't know if you asked the wrong question, or I give the wrong answer :-)

    It is always good to attach screenshots, background to the issues, and what exactly you are trying to solve.

    You wrote:

    So to summarize:

    -ESXi uses consume to decide when to start the recovery techniques of memory RAM

    -To rightsize VMs for RAM, you must use the counter inside the BONE, not active

    -If you do not want to commit too much RAM and have a lot of the balloon you should right-size your virtual machines or extend the RAM / buy additional servers

    Please correct me if I'm wrong.

    My answer: you are right. If you do not have access to the guest operating system, then Active VM and VM consumed are better guess. Better than nothing :-)

    Sizing sizing of VM and ESXi are 2 different things (2 use cases) different. Please don't mix up the as soon as they are confusing. Do not try to resize the virtual machine when you are sizing the ESXi, unless you have a control total on them.

    You wrote:

    The question for me is: how many cache (standby under Windows) inside the VM is really necessary for the virtual machine to get the best performance without wasting resources. Perhaps this issue can be resolved only by tests or do you know another option?

    My answer: it's a different topic brother :-) It is different, because it is within a guest operating system (Windows in this case) and is no longer at the level of the ESXi. Suggest that post you another, possibly under Windows or vSphere too :-)

  • SQL plan

    Hello
    on the 11g R2,.
    -Each sql plan is unique in a database? If Yes, how is identified and how to find it when you query SQL_ID?
    -Any request to find a DBA_HIST_SQL_PLAN sql plan?
    My problem is:
    I ran a long query and when running I was looking at V$ session to have SQL_ID and SQL_ADDRESS. Today, when I run the following qury, no row is returned. Changed address between the two?
    SELECT DECODE(id,0,'',
    LPAD(' ',2*(level-1))||level||'.'||position)||' '||operation||
    DECODE(id,0,'',' Option: ')||options||
    DECODE(id,0,'',' Object Name: ')||DECODE(id,0,'',NVL(object_name, 'None'))||
    DECODE(id,0,'',' Object Number: ')||DECODE(id,0,'',NVL(TO_CHAR(object#), 'None' ))|| 
    DECODE(id,0,'Cost = '||position) 
    SQL_Query_plan 
    FROM v$sql_plan
    CONNECT BY prior id = parent_id
    AND address = upper('000007FF38329428')
    START WITH id =0 AND address = upper('000007FF38329428');
    Thank you.

    Several plans like Justin mentioned.

    For the map, is it not easier to do all this?

    SELECT * FROM TABLE(DBMS_XPLAN.DISPLAY_AWR('',''));
    
  • SQL Plan Management vs. cursors

    Hello

    We have recently implemented SQL Plan Management in our Production environment after two weeks of testing in development.

    Wouldn't you know, we proceeded to six questions which lasted much longer (, instead of less than a minute).

    The common denominator of these queries is that they are used as part of a loop of slider and make the use of bind variables. All other queries (and we talk about hundreds) was fine. The data has not changed for relevant queries.

    Is there a 'problem' with SPM and sliders/bind variable that anyone could have heard?

    Mark

    When you say SQL Plan Management, do you mean Plan SQL schedules?

    Baselines work by locking in a sql statement to a particular regime.

    The only problem that come to mind is that the referenced plan is not suited for all bind variable values.

    The first thing to look at would be the execution plans and they are in the base line, etc. - see DBMS_XPLAN. DISPLAY_AWR, also cross references to ash.
    Keep an eye on other features such as feedback of the cardinality.

    On 11.2.0.2, one thing to watch is a big SPM and recursive merge instructions involved. There is a patch, but it's usually pretty obvious if you are affected by this bug, for example
    http://OraStory.WordPress.com/2011/03/01/SQL-baseline-capture-BU/

Maybe you are looking for