Behavior of ctx_ddl.optimize_index in FULL mode

Hi,
FULL mode is documented to "pick up where it left off".
We understand that the way you can "tell" where the optimize_index process is is by looking at:
CTXSYS.dr$index.idx_opt_token
If the optimize has stopped and you run it again, it will "continue" from that token.
We were wondering:
We want to "trick" the optimize program so that it starts from a different token.
So we were thinking we could update that idx_opt_token value (while the optimize program isnt running) to the value that we want.
Has anyone ever done this before? We tried it in a development environment and it seems to work, but would prefer to know that we are not breaking something.
Need to do this as opposed to optimizing in token mode and writing our own wrapper to do things token by token.
The reason is because when we run in token mode or any mode, there is some initial query against the $n table that takes a long time and we dont want to keep doing that again and again.

As a general rule, you should never update any data dictionary directly. There may be unknown cascading effects and Oracle does not offer support once you have modified a data dictionary directly.

Similar Messages

Optimizing in full mode does not removes false hits any more from COUNT_HIT

Hello,
I am using the COUNT_HITS fonction to get the number of hits for a full text query with the EXACT parameter set to false or true.
If I am not wrong the ctx_ddl.optimize_index('MY_INDEX','FULL') is suppose to remove false hits from COUNT_HITS where the EXACT parameter is set to false but it does not work any more. It still takes into accound records that have been deleted.
Is there other actions to do to remove false hits from COUNT_HITS?
Thank you for your help.
Kind regards,
Fred

any way to make the pinned site shortcut (not all sites) open in full screen mode when double clicked?
I think it does which is a surprise. E.g. if you last closed an IE window in Fullscreen mode it can open up again in Fullscreen mode. This may be in response to user complaints about Maximized mode which would still be a problem because that
is the OS doing. However, even there it has always been possible to open a window from a shortcut in Maximized mode. The problem is that any windows spawned from that one will have the original window's Normal mode attributes in terms of size and
shape.
FYI
Robert Aldwinckle

Ctx_ddl.optimize_index in dbms_job

I am trying to set up the jobs to optimize our various interMedia indexes at night, but am having an issue with ownership. The index exists in the user_a schema. As CTXSYS, I can successfully:
ctx_ddl.optimize_index('USER_A.INDEX_NAME', 'full');even with 'set role none' in force.
But if I submit this as a job under CTXSYS, the job fails (putting an entry in the alert log) with a DRG-10016 (I'm pretty sure this was the error number) saying that CTXSYS needs to be the owner of the object to modify the object.
Why can I do the ctx_ddl.optimize_index at a SQL*Plus prompt with 'set role none' and not through a job?
Also, is it correct to set up this job as CTXSYS instead of USER_A? USER_A cannot even execute CTX_DDL in a job without a direct grant - and I haven't seen any examples saying this is the way to do it.

As a general rule, you should never update any data dictionary directly. There may be unknown cascading effects and Oracle does not offer support once you have modified a data dictionary directly.

Images in Aperture appears oversaturated in the Full mode. Please help?

I'm using 30" Apple HD screen, and when I opening photo in the normal mode all the colors are fine. However once I change it to full screen mode the Red and Yellow become overstaurated. Please help!

Are you using two displays? If you are then odds are your other display (not the 30") is set as the default display. When this happens the full mode uses the other display's profile (the default) in full screen mode.
What you need to do to correct it is to go into the color sync utility. Under the "Devices" tab go down to "Displays". Toggle it show all your displays. Select your 30" ACD and then to the right select the "Make Default Display" button.
Jon Roemer
site: http://www.jonroemer.com/
blog: http://jonroemer.typepad.com/jon_roemer/

I cant watch youtube on full mode its always hangs only in firefox

i cant watch youtube on full mode its always hangs only in firefox

Start Firefox in Safe Mode to check if one of the extensions (Firefox/Tools > Add-ons > Extensions) or if hardware acceleration is causing the problem (switch to the DEFAULT theme: Firefox/Tools > Add-ons > Appearance).
<b> To Enable SafeMode </b>
*You can open Firefox 4.0+ in Safe Mode by holding the '''Shift''' key when you open the Firefox desktop or Start menu shortcut.
* Or open the Help menu and click on the '''Restart with Add-ons Disabled...''' menu item while Firefox is running.
*''Once you get the pop-up, just select "'Start in Safe Mode"''
If it works in Safe Mode and in normal mode with all extensions (Tools > Add-ons > Extensions) disabled then try to find which extension is causing it by enabling one extension at a time until the problem reappears.
Close and restart Firefox after each change via "Firefox > Exit" (Windows: Firefox/File > Exit; Mac: "Firefox > Quit Firefox"; Linux: "Firefox/File > Quit")
* https://support.mozilla.org/en-US/kb/troubleshoot-firefox-issues-using-safe-mode
*https://support.mozilla.org/kb/Safe+Mode
*https://support.mozilla.org/kb/Troubleshooting+extensions+and+themes

Payroll extractor 0HR_PY_1 do not return any data in RSA3 in full mode

Hi,
I have activated Payroll extractor 0HR_PY_1.
I use RSA3 to check the extractor.
It does not return any data(0 records) in full mode.
Do i need to do activate any other object, apart from the datasource?
Thanks in advance for your help.
Regards,
Suresh.

Hi Suresh,
HR extractors need additional authorizations. When the extractor is run, there would be a check if the user is authorized to extract data. Check SAP notes 672514 & 397208. Also find my previous post on this.
Re: Extractor 0HR_PY_1 does not return data to BW
Best Regards,
Murali.

*virtualbox gnome3* Cpu is too high 99% when using gnome3 full mode

1G ram, 128 video ram. when click the 3d item, cpu rise to 99%..cannot use full mode.system hanged

Am i reading right you are using Arch with Gnome3 in virtualbox if so you are using much to much video memory 60mb is about right for vbox. 3d is experimental and more than likely will not work with gnome3.

Flickring screen in Full mode my mac jammed up!

I have been an ardent user of aperture for years now and got used to the great workflow. Ever since version 1 I have been enjoying working with aperture. It is the workflow that kept me from changing.
I got a new MacPro 12 GB ram, 4TB HDD my library mostly indexed files sits on a raided 2TB for maximum of comfort. I did this in an effort to keep me from writing the posts I writing lately. I do not have the time rally to do this but reality is something different.
Today I am working on a job I need to finish quickly and in aperture workflow manner. So I switched to full screeeeeeeeeeeeeeeeen mode and after pressing 'H' for the HUD my screen started flickering. please look at the video I posted in my flickr account: http://www.flickr.com/photos/mawas/4719742585/
Now this is the second time this happened in a short time after I got a new mac and upgraded to 10.6.4! After the screen had jammed up : http://gallery.me.com/masuwa/100103 I needed to hard boot my mac I am getting nervous!

Hello, I also saw you post about this over on Flickr. I'll post my experiences here as well. Yes, I've had the screen flickering and generally going haywire and bringing down the whole system. Had to restart from the power button.
As with your experience, this has all happened since the update to 10.6.4. Frequent hangs as well. I've just repaired my main work database for Aperture. We'll see whether that helps any.

Selection of Requests in PSA while loading DTP in Full-Mode

Hi all,
i'm using the DataSource 0uc_sales_stats_01 to load data into BW 7.0.
The InfoPackage is manually scheduled.
In older posts, i did not find any solutions.
Problem description:
i can't delete old requests in PSA, because they are needed by other users.
the DataSource is connected with an DSO via transformation.
in the created DTP, i can only choose full extraction mode (delta mode is not supported for this DataSource).
in my workaround i would like to select only the new requests to avoid duplicate loads into my DSO.
but in the endroutine or startroutine, i can not access the PSA request numbers.
also, there is no possibility to select requests in DTP filter...
does anyone know a solution for my problem?
kind regards,
Philipp

hi Philipp,
Its just a thought, just pretend that the old requests are successfully updated to the dso . Next time dtp will not pick those request. If you can find a way to update tables like (rsodsactreq ) with old request and success status,then you can skip those requests next time.
cheers

Can xmobar show the # of windows open in full mode?

Hi, I'm a dwm user trying xmonad. Since I'm on a netbook, I use monocle mode a lot, and I thought it was really useful how it showed me the number of windows that were open in the bar. xmonad does not do this, and I was wondering if there's a way to make it do so.
Thanks.
Last edited by thursdayknights (2009-12-16 19:52:20)

Take a look at this extension. <br />
https://addons.mozilla.org/en-US/firefox/addon/r-kiosk/
Real Kiosk is a Firefox extension that defaults to full screen ...

Custom part made in evaluation mode transfers to full mode when license bought?

I'm just wondering if all the custom parts I made while evaluating Multisim 11 will transfer over when I actually apply my license. Just to be on the safe side before I make the switch
Solved!
Go to Solution.

Hi,
You won't have any problems, all custom parts you made will be in the User database.
Tien P.
National Instruments

Volume slows down even if the volume on full mode

The caller can't hear me but i can hear them, i tried cleaning the headset jacket as well as the phne speaker with soft brush but still not working, i tried resetting the network but still the same. I tried restoring it , it was ok after the restoration but the problem comes back after one call received. I can't call the support coz i dont have any apple care and it is costly to call phone support coz as you know we need a credit card and i don't have one. The service center is far away form us and its too expensive. Please advice me further maybe you still have solutions, new solutions that was not shared here in the Apple Support Community. Please hoping i can hear from you then, i really love my iphone and for me if its like me getting sick if my iphone is not working well consider my financial status now...Please help me. Thank you an more power to us, iphone users...

Try:
- Reset the iOS device. Nothing will be lost
Reset iOS device: Hold down the On/Off button and the Home button at the same time for at
least ten seconds, until the Apple logo appears.
- Reset all settings
Go to Settings > General > Reset and tap Reset All Settings.
All your preferences and settings are reset. Information (such as contacts and calendars) and media (such as songs and videos) aren’t affected.
- Restore from backup. See:
iOS: Back up and restore your iOS device with iCloud or iTunes
- Restore to factory settings/new iOS device.
If still problem, make an appointment at the Genius Bar of an Apple store since it appears you have a hardware problem.
Apple Retail Store - Genius Bar

Oracle text related internal procedure taking a lot of time in our Production database

Hi,
I am on Oracle 11.2.0.2 on Linux. I have Oracle Text implemented in all my databases for fuzzy search. I am seeing the following Oracle TExt specific internal procedure to be among the Top SQL in my AWR in production. This is during business time.
SQL ordered by Elapsed Time
Resources reported for PL/SQL code includes the resources used by
all SQL statements called by the code.
% Total DB Time is the Elapsed Time of the SQL statement divided
into the Total Database Time multiplied by 100
%Total - Elapsed Time as a percentage of Total DB time
%CPU - CPU Time as a percentage of Elapsed Time
%IO - User I/O Time as a percentage of Elapsed Time
Captured SQL account for 59.3% of Total DB Time (s): 120,379
Captured PL/SQL account for 33.8% of Total DB Time (s): 120,379
Elapsed Time (s)
Executions
Elapsed Time per Exec (s)
%Total
%CPU
%IO
SQL Id
SQL Module
SQL Text
23,476.22
205,095
0.11
19.50
16.21
7.88
ddr8uck5s5kp3
begin ctxsys.drvdml.com_sync_i...
Note that the sql id ddr8uck5s5kp3 has this sql:
begin ctxsys.drvdml.com_sync_index(:idxname, :idxmem, :partname); end;
Also note that I have the procedure to optimize the indexes (ctx_ddl.optimize_index in FULL mode) set up every night to run at 3 am for all our Oracle Text indexes. Is there anything else needed. I don't know why the procedure I showed above in the AWR report takes so much time and why it is among our Top sql.
I will be very thankful for guidance in this regard.
Thanks,
OrauserN

This is the internal call which drives the SYNC call for a text index. Effectively all the indexing of new and updated data in your text-indexed table is contained within this call.
If you're using parallel SYNC you will see this call contained with a SELECT query - that select is executed as a parallel query on a table function, which is the way we divide up the work between parallel slaves.

Oracle Text USER_LEXERについて

Oracle Text を USER_LEXER で使用したいと考えております。
後述のサイトを参考にプロシージャをcreateしようとしたのですが
以下のエラーになってしまい実行出来ませんでした。
対処方法ご存じでしたら教えて頂けないでしょうか。
＠エラー内容
PROCEDURE QUERY_OTのエラーです。
LINE/COL ERROR
0/0 PL/SQL: Compilation unit analysis terminated
4/20 PLS-00904:
オブジェクト'PUBLIC.CTX_ULEXER'にアクセスするには権限が不十分です
＠実行したSQL文
CREATE OR REPLACE PROCEDURE query_ot
p_target IN VARCHAR2,
p_tab IN CTX_ULEXER.WILDCARD_TAB,
p_result IN OUT VARCHAR2
AUTHID CURRENT_USER
IS
cursor cur_oti is select token from oti;
BEGIN
p_result := '<TOKENS>';
for now in cur_oti loop
if instr(p_target, now.token) > 0 then
p_result := p_result || '<WORD>' || now.token || '</WORD>';
end if;
end loop;
p_result := p_result || '</TOKENS>';
END query_ot;
＠参考サイト
http://www.nacky.info/wiki/index.php?OracleText
＠環境
OS: Ubuntu 12.04LTS
Oracle: 11.2.0-1.0 (64ビット版)
Edited by: 955082 on 2012/08/26 20:42

ご回答ありがとうございます。
FORMAT COLUMNとUSER_DATASTOREを試したところ、
当方の望みの動作を確認出来ました。
（マテリアライズド・ビューは後に確認させて頂きます）
ただ、検索対象のデータを更新したの索引の同期が取れず
また問題につまずいてしまいました。
度々で恐れ入りますが、同期の方法ももしよろしければご教授頂けないでしょうか。
USER_DATASTOREを実装したSQLと確認手順は以下の通りです。
＠データやプロシージャ作成SQL
--検索対象テーブル
drop table ot;
create table ot
( id number primary key,
text varchar2(80),
del_flg number,
type varchar2(80) );
insert into ot ( id, text, del_flg, type ) values ( 1, 'The cat sat on the mat', 0, 'TEXT' );
insert into ot ( id, text, del_flg, type ) values ( 2, 'The dog barked like a dog', 0, 'TEXT' );
insert into ot ( id, text, del_flg, type ) values ( 3, '日本オラクル株式会社', 0, 'IGNORE' );
insert into ot ( id, text, del_flg, type ) values ( 4, '全日本自動車協会', 1, 'TEXT' );
commit;
-- 辞書テーブル
drop table oti;
create table oti
( token varchar2(80) );
insert into oti ( token ) values ( '日本' );
insert into oti ( token ) values ( 'オラクル' );
insert into oti ( token ) values ( '日本オラクル' );
insert into oti ( token ) values ( '自動車協会' );
insert into oti ( token ) values ( 'Ora' );
commit;
-- 索引作成プロシージャ
CREATE OR REPLACE PROCEDURE ind_ngram
v_a IN VARCHAR2,
v_b IN OUT VARCHAR2,
v_c IN BOOLEAN
IS
cursor cur_oti is select token from oti;
BEGIN
v_b := '<tokens>';
for now in cur_oti loop
if instr(v_a, now.token) > 0 then
v_b := v_b || '<word>' || now.token || '</word>';
end if;
end loop;
v_b := v_b || '</tokens>';
END ind_ngram;
--索引検索プロシージャ
CREATE OR REPLACE PROCEDURE que_ngram
v_a IN VARCHAR2,
v_b IN CTX_ULEXER.WILDCARD_TAB,
v_c IN OUT VARCHAR2
IS
cursor cur_oti is select token from oti;
BEGIN
v_c := '<tokens>';
for now in cur_oti loop
if instr(v_a, now.token) > 0 then
v_c := v_c || '<word>' || now.token || '</word>';
end if;
end loop;
v_c := v_c || '</tokens>';
END que_ngram;
--データストアプロシージャ
create or replace procedure myproc(rid in rowid, ret in out nocopy varchar2) is
begin
ret := null;
for c1 in (select text from ot
where rowid = rid and del_flg = 0)
loop
ret := c1.text;
end loop;
end;
--CTX_DDL
BEGIN
CTX_DDL.drop_PREFERENCE('my_lexer');
CTX_DDL.CREATE_PREFERENCE('my_lexer', 'user_lexer');
CTX_DDL.SET_ATTRIBUTE('my_lexer', 'index_procedure', 'ind_ngram');
CTX_DDL.SET_ATTRIBUTE('my_lexer', 'input_type', 'varchar2');
CTX_DDL.SET_ATTRIBUTE('my_lexer', 'query_procedure', 'que_ngram');
END;
BEGIN
CTX_DDL.drop_STOPLIST('my_stoplist');
CTX_DDL.CREATE_STOPLIST('my_stoplist', 'basic_stoplist');
END;
BEGIN
ctx_ddl.drop_preference('my_datastore');
ctx_ddl.create_preference('my_datastore', 'user_datastore');
ctx_ddl.set_attribute('my_datastore', 'procedure', 'myproc');
ctx_ddl.set_attribute('my_datastore', 'output_type', 'varchar2');
END;
--索引作成
CREATE INDEX a ON ot(text)
INDEXTYPE IS ctxsys.context
PARAMETERS
LEXER my_lexer
STOPLIST my_stoplist
datastore my_datastore
＠確認手順１：索引を表示
# USER_DATASTORE実装前はid=3とid=4のデータが索引に格納されていたが、
# id=4のdel_flgを1に設定してINSERTしたためid=3 のデータのみが格納されている(正常)
SQL> select token_text from dr$a$i;
TOKEN_TEXT
オラクル
日本
日本オラク
＠確認手順２：id=3のdel_flgをオン
SQL> update ot set del_flg=1 where id=3;
1行が更新されました
SQL> commit;
コミットが完了しました。
＠確認手順３：同期コマンドを実行
SQL> exec CTX_DDL.SYNC_INDEX('a');
PL/SQLプロシージャが正常に完了しました。
SQL> exec CTX_DDL.OPTIMIZE_INDEX('a','full');
PL/SQLプロシージャが正常に完了しました。
＠確認手順４：再び索引を表示
# id=3のdel_flgに1を立てていたため索引はでないはず・・・だが表示されてしまった。更新されていない模様(異常)
SQL> select token_text from dr$a$i;
TOKEN_TEXT
オラクル
日本
日本オラクル
---

$I rows: 24,973,528 takes 8 hours to optimize against 0% fragmentation

We are using UCM 10g for our Web Content Management environment. We use Oracle Text Search (DB version 11.1.0.7) as our search engine. To maintain performance of the indexes, we have a daily job scheduled in database to optimize the indexes. But it takes around 8 hours to optimize a 0% fragmented index. Any idea why it takes long time? Any suggestions to reduce the time it takes to run?
Before running optimization script - Fragmentation details:*
total size of $I data: 838,699,490 (799.85 MB)
$I rows: 24,973,528
estimated $I rows if optimal: 24,973,528
estimated row fragmentation: 0 %
Time taken to optimize the above index:_
=============== FT_IDCTEXT2 Start Date and Time: 28=FEB-2013 02:37:22===========
PL/SQL procedure successfully completed.
=============== FT_IDCTEXT2 End Date and Time: 28=FEB-2013 10:55:55=============
Script:*
select '=============== FT_IDCTEXT2 Start Date and Time: ' ||sysdate||'=============== ' from dual;
begin
ctx_ddl.optimize_index('FT_IDCTEXT2','FULL', parallel_degree =>'10');
end;
select '=============== FT_IDCTEXT2 End Date and Time: ' ||sysdate||'=============== ' from dual;

Hi ,
I believe it would be a question for DB to tell why the fragmentation is taking such a long time for un-fragmented db schema . For UCM to improve it's search , indexing processes this action plan is done . But , since the DB itself is taking such an inordinate long time to do optimization , I think it would be a DBA question :-) .
Thanks,
Srinath

Behavior of ctx_ddl.optimize_index in FULL mode

Similar Messages

Maybe you are looking for