Extract(Read) > Transform(Test&Apply) > Load(Learn) Blog

December 19, 2010

Oracle DW Global Leaders Program for EMEA – 2010

Filed under: Oracle Seminars — H.Tonguç Yılmaz @ 3:51 pm

I attended the 5th official Oracle EMEA DW Global Leaders meeting last week hosted by Oracle DWH Product Management at Barcelona. Each year we meet twice, last meeting was at Krakow Poland at June. I believe what Oracle does with this program is very unique, each time we as the largest Oracle DW customers at EMEA meet with the top management of Oracle DW prodcut management and each other to exchange information. Program has now 65 participants from 13 different countries. Oracle defines this program as: “The DW global leaders program is a strategic Customer Program within Oracle DW Development to ensure successful and referenceable DW Clients and support our future DW product directions through ongoing, close and proactive relationship. In addition we want to offer a platform for our members to share experience and best practice in Data Warehousing”.

Additional to these meetings twice a year there are WebSeminar Series by customers monthly, Customer Advisory Boards where we discuss on the future developments of Oracle DW solutions, Conf-Calls on specific DWH topics with other DW Global Leaders customers, escalation of SRs and establishing relationship with DWH development, DWH architecture review and healthchecks and an assigned DWH Product Manager as contact point for DW Global Leaders. At Barcelona my company Turkcell Technology was announced as the Communications Associate Partner of Oracle DWGL program and we presented our Data Quality Management platform we developed for Turkcell ICT DW needs during 2010. Next meeting will be at Athens June 2011. If you want to know more about DWGL program or our DQ Management platform you can drop me an email.

Advertisements

August 20, 2010

TROUG: hayal mi yoksa gerçekleşmesi için yeterli olgunluk düzeyine ulaştık mı?

Filed under: Oracle Events,Oracle Other,Oracle Seminars,Other,TROUG — H.Tonguç Yılmaz @ 7:09 am

Konu paralelinde bu alandaki ilk Türkçe yazımı paylaşmak istedim, Mayıs 1999’da kurulan OracleTurk yahoo grubu forumunda uzun senelerdir yöneticilik yapıyorum. Grup özellikle 2002 yılı ile birlikte hareketlenmeye başladı ve 2007 ile 2008 yılları tartışmaların en hareketli olduğu seneler oldu, bu dönemde gruba wordpress üzerinde bir blog da kurduk. Sayısını şu anda hatırlayamadığım onlarca ücretsiz seminer yaptık grup üzerinden, çoğu Üniversite’lerde gerçekleşen bu seminerler Oracle veritabanı kullanıcısı olarak kariyerine yeni başlayacak başarılı gençlerle buluşmamızı sağladı, bugün itibari ile grubun 1800 üzerinde üyeliği bulunmakta.

Bu kısa tarihsel girizgah sonrasında tartışmak istediğim konuya yavaş yavaş girmek istiyorum, “Bağımsız Türk Oracle Kullanıcıları Grubu” benim 2004 yazından bu yana OracleTurk başta olmak üzere çeşitli platformlarda dile getirdiğim ve çabasında olduğum bir konu oldu. Özellikle yurtdışında Slovenya, Estonya gibi Ankara’dan daha küçük denebilecek ülkelerin User Group aktivitelerini gördükçe önemli potansiyelinden her fırsatta bahsedilen ülkemizin kullanıcı grubunun olmadığı gerçeğinden hep utandım sohbetlerde. Fakat zamanla bu grubun olgunlaşamamasının arkasında Türkiye’nin sosyal dinamiklerinin olduğunun da farkına varmaya başladım ve zamana bırakmayı kabullendim.

Benzerlik kurmaya çalışırsak Kuzey Avrupa’da insanlar oy verirken belki çevreci politikalar onlar için çok önemli olabilir ama Türkiye’de durum farklı biliyoruz. Tanıdığım bilgili/tecrübeli bir çok uzmanın hemen hemen hiç paylaşımda bulunmamasına ve bu gibi konularda çaba göstermemesine, destek önerdiğimde de “eve gidip dinleniyorum zar zor, benim zamanım yok bu işlere ayıracak.” “Girmişsin bir işe piyasadaki insan kalitesini bilmeden konuşuyorsun, bu platformlarda canın çok sıkılır saçma sapan talepleri ile insanların uğraşmaktan.” benzeri cevaplarına o kadar alıştım ki, son senelerde artık cidden sesim çıkmaz olmuştu. Bende başarılı gençler ile zaman geçirmeye ve onlara kariyerlerin başında yardımcı olmaya odaklanmaya karar verdim.

Bazı şeyleri zamana bırakırken en güvendiğim referansım eğer cidden bir ihtiyaç var ise bu ülkede bir Oracle kullanıcı grubuna elbette kaliteli gençlerin arkadan yetişmesi ve 1975 sonrası doğan teknolojiye ilgili kişilerin yönetim kademelerinde ilerlemesi ile kullanıcı grubu fikri yavaşta olsa olgunlaşacak ve sonuçlanacaktır düşüncesi idi. Son seneler içinde WordPress, Blogger, Facebook, Linkedin gibi paylaşım platformlarının yaygınlaşması ile kaliteli paylaşım peşinde olan bir çok Türk blogger, forum takipçilerinin bu ortamlarda aktivitelerinin arttığını gözlemledim, içten içe hep mutlu oldum. Bugün kendi RSS okuyucumda Türk blogger’lar için özel bir sekme tutuyorum, bir yeni giriş düştüğünde buraya heyecanla açıyor ve hemen okuyorum :)

Dün Oracle Türkiye ofisinde bir toplantıya katıldım, toplantıda Türkiye Oracle ofisi satış ekibi desteği ile TROUG oluşumunun harekete geçirilmesi konuşuldu. Lokalde satış ekibinde önemli değişiklikler olmuş, bu toplantı daha önceki deneyimlerimden çok daha heyecan verici ve sonuç odaklı idi, mutlu oldum ve destek vereceğimi garanti ettim elbette. Daha önceki deneyimlerime hiç girmeyeceğim bu yazıda, çok da gerekli değil cidden. Bu noktada bence biraz da neden TROUG sorusuna yanıt aramakta fayda var. Beni en çok etkileyen kullanıcı grubu bu anlamda UKOUG olmuştur hep, her yıl Aralık aynda gerçekleşen konferanslarına iki kez katılma şansım oldu, bu kadar kaliteli bir içerik başka hiçbir yerde tecrübe etmedim diyebilirim. Dolayısı ile ilk önemli motivasyon bence kaliteli paylaşım ve özel ilgi grupları yaratarak odaklanma ile kullanıcı grubu aktif üyelerin uzmanlığını arttırmak, bilgiyi paylaşarak birlikte gelişmek. Bu paylaşım blog ve forum’lardakine göre çok daha derin zira gönüllü sunucular için sunuma hazırlık bir anlamda ciddi ders çalışma gerektirecektir, tanışma-tartışma fırsatı ve birlikte zaman geçirerek yakınlaşma-dostluklar üyeler arasında ilişkileri pekiştirecektir.

İkinci motivasyon tabii ki birlikten doğacak güç olacak, birkaç bin kişilik bir kullanıcı grubu sponsor bulmakta güçlük çekmeyecektir, bir çok ücretsiz seminer ve yurt dışından uzman kişileri davet etmek mümkün olabilir. Oracle ile konuşurken daha güçlü tek bir ses olarak taleplerde bulunabiliriz. Ayrıca piyasada iş imkanları bu kullanıcı grubu üzerinden sonuçlanabilir, üyelere iş fırsatlarında grup aktivitesi önemli bir referans oluşturabilir.

Peki gelelim tüm bunları yazıya dökmemin nedeni olan tetiklemek istediğim tartışma sorularına, sizce;
– böyle bir oluşum neden faydalı olur, istekten ziyade cidden ihtiyaç hissediyor musunuz? yoksa mobil internet ile güçlenen internet üzerindeki paylaşımların takibi yeterli tatmini yaratmaya devam eder mi?
– eğer varsa benzer olumlu ya da olumsuz kişisel deneyimleriniz nedir?
– böyle bir oluşum içinde mutlaka olmalı dediğiniz kişiler/gruplar kimler(Oracle Forum Türkiye isimli grubun yöneticisi Zekeriya Bey de dün toplantıda idi, onun da çok istekli ve destek olmaya hazır olduğunu hissettim), siz destek olmak için gönüllü olmak ister misiniz? Yanıtınız evet ise blog adresinizi yorum olarak belirtir misiniz.

Son bir not “Oracle kullanıcıları grubu” derken artık çok geniş bir topluluktan bahsediyoruz, muhtemel özel ilgi grubu örnekleri vermek gerekirse: Veritabanı geliştirme(SQL, PL/SQL, Apex, Java, Jdev&ADF, Forms&Reports, Windows&.Net, XML, BI&DWH ÖİG gibi), Veritabanı yönetimi(Yüksek erişebilinirlik&Grid Computing, Enterprise Manager&Grid Control, Kurulum&yükseltme ÖİG gibi), Genel(Eğitim&Sertifikasyon ÖİG gibi), Fusion Middleware(Weblogic, SOA ÖİG gibi), Mimari(Solaris, Linux, VM ÖİG gibi), Paket uygulamalar(E-business Suite, Siebel, Peoplesoft, JD Edwards, Primavera ÖİG gibi).

May 22, 2010

Oracle Partitioning and Parallel Execution for Performance seminar by Tanel Poder

Filed under: Oracle 11g New Features,Oracle Business Intelligence,Oracle Seminars — H.Tonguç Yılmaz @ 2:44 pm

Tanel was our visitor last week as I mentioned earlier, he did a one day seminar on partitioning and parallel execution. There were lots of interesting discussions for me, I tried to share some of my notes just to give an idea how Tanel’s experience may affect your daily Oracle database administration or development routine.

Other ways for detecting partition pruning rather than Pstart and Pstop info at DBMS_XPLAN:
– Event 10128, semi documented in Note 209070.1, create partition pruning tracing table under your schema KPAP_PRUNING
– Samping ROW_WAIT_OBJ# on V$SESSION with Snapper or ASH, you get the dba_object.data_object_id values
– Taking snapshots of V$SEGMENT_STATISTICS logical IO for each partition segment

Partitioning for OLTP DML Concurrency
– Partitioning can be effectively used for reducing concurrency on right hand index blocks in OLTP(ever increasing sequence or timestamp based columns)
– Hash sub partitioning for meaningless pkey columns, the kgghash() function used for row placement by hash is optimized for power of 2 buckets
– Partitioning for data load concurrency, below option will take TM lock only on the partition and allow multiple load operations on the same table
INSERT /*+ APPEND */ INTO taget_table PARTITION(target_partition)
SELECT * FROM source_table;

Parallel Query Execution change with 10g
– In 9i the QC unparses transformed query text and send a copy for each slave for separate execution. In 10g+ each slave executes the same cursor

Parallel execution more resources
– Especially because of the PX communication overhead
– And additionally PX operation on a table requires a segment level checkpoint.
– With 11.2 in-memory parallel execution if PARALLEL_DEGREE_POLICY is set to AUTO
– So not everything should use paralelism and not every system is a candidate for parallelism, even in DWs(like ETL metadata updates etc.)

For parallel DML always remember to enable parallel dml at session level
– alter session enable parallel DML;
insert /*+ parallel(4) pq_distribute(eo none) */ into etl_orders eo
select * from tab;
– Query PDML_ENABLED, PDML_STATUS, PDDL_STATUS, PQ_STATUS columns of V$SESSION to check parallel DML is enabled or not for a session
– if data read is highly skewed, different slave produce way amount of rows, you may want to distribute rows for loading with pq_distribute(eo random) hint

Major Parallel Execution bottlenecks
– Serially executed steps in PQ plans, check for P->S and S->P operations in the middle of the execution plans. “IN-OUT” column is empty for serial operations in DBMS_XPLAN
– Too much communication between slaves, if possible replace a BROADCAST distribution with HASH or PART
– Skewed row distribution between producers and consumers, check V$PQ_TQSTAT if some slaves produced/consumed signaficantly more rows than others in the slave set, swith form HASH to BROADCAST ot PARTITION distribution
– Unbalanced hardware configuration, not enough IO bandwith, the very common reason, query ASH or run snapper of PX slaves to measure time waited for IO

Parallel Execution waith events and sql trace
– Metalink Note 191103.1 for PQ wait events
– After 10g+ before tkprof utulity to aggregate the multiple tracefiles produced for each slave use trcsess utulity

Parallel Execution and RAC
– In 10g set PARALLEL_INSTANCE_GROUPS parameter in each instance to assign the instances to groups
ALTER SYSTEM SET instance_groups = ‘rac01′,’all’,’batch’ ;
ALTER SYSTEM SET instance_groups = ‘rac02′,’all’,’batch’ ;
ALTER SYSTEM SET instance_groups = ‘rac03′,’all’,’oltp’ ;
ALTER SYSTEM SET instance_groups = ‘rac04′,’all’,’oltp’ ;
— Before running the PQ statement SET parallel_instance_groups
ALTER SESSION SET parallel_instance_groups = ‘batch’ ;
— The slaves will run only on the instances associated with the ‘batch’ group
– In 11g again parallel_instance_groups controls the allowed groups where QC can schedule inter instance PQ slaves, but this time slaves can only run on the instances which allow running the same “service” the QC connected through.
— Alternatively
ALTER SESSION set PARALLEL_INSTANCE_GROUP = ‘service1′,’service3’ ;

Parallel Execution Overview
– PX COORDINATOR: Query Coordinator(QC), distributes the work to slaves and returns results back to the client. Additionally performans work if something has to be done serially.
– PX SEND QC: Sends produced rows to QC via a table queue
– PX BLOCK ITERATOR: Operation which picks different ranges ıf PX granules to read different slaves
– Table Queues: In-memory communication channel between PX slaves, if SGA_TARGET is in use or PARALLEL_AUTOMATIC_TUNING is set to TRUE a memory buffer inside large pool
– QC(RAND): The row source is sending its rows back to QC

Things to remember
– Watch out for accidently enabled paralelism especially on OLTP, after parallel index builds and rebuilds remember to ALTER INDEX NOPARALLEL
– The Cost we see in DBMS_XPLAN does not include PX distribution cost, however CBO does choose PC distribution based on cost(10053 CBO tracefile shows that the costing is done)
– Use PARALLEL_MIN_PERCENT parameter to decide the minimum % of the requested Degree of Paralellism(DOP) query should be able to allocate in order to run the query
– PARALLEL_EXECUTION_MESSAGE_SIZE default to 16KB after 11.2
– Use ALTER SESSION FORCE PARALLEL QUERY PARALLEL n; to force session level paralelism

Oracle Partitioning and Parallel Execution for Performance seminar by Tanel Poder

ps: Check out Oracle Sweden User Group(ORCAN) Event presentations of Hüsnü

June 11, 2009

Oracle 11g Data Warehousing Masterclass with Mark Rittman

Filed under: Oracle Seminars — H.Tonguç Yılmaz @ 11:55 am

Mark Rittman will be in Istanbul for a two days seminar this month(30 June-1 July), most probably I will be there and here is the details of this seminar.

February 5, 2009

What can be the purpose with “DBA 2.0” synonym by Oracle?

Filed under: Oracle Seminars — H.Tonguç Yılmaz @ 2:30 pm

Today Mr.Mughees A. Minhas, director of product development/database management was our guest at my company and he gave a presentation on “DBA 2.0” idea. The presentation had three parts:
– Subjects related to how to sustain optimal performance with Oracle database,
– Subjects related to preserve order amid change with Oracle database,
– Subjects related to managing more with less with Oracle database

Lots of nice powerpoint slides related to extra cost options were discussed as usual :) One of my comments after the presentation was Oracle DBAs’ lives are getting much more complicated day by day and what Oracle marketing stuff is forcing IT managers to believe is completely the opposite, this is not fair just to sell more.

Change is of course inevitable and we are all somehow have to accept the changes which are managed by the guys at San Francisco head quarters, but in my opinion forcing management that if they migrate their DBA stuff to version 2.0 somehow everything will be easier and the stuff will be doing absolutely less is not accurate. For example with RAT DBA’s are now also testers, with ASM DBA’s are now storage admins right? Until now Oracle was able to automate easier parts like Memory, Undo, Segment Space management and they aggressively continue to come up with new features which need manual attention. But calling the older DBAs who are using Statspack not Grid Control or file system backup not RMAN as version 1.0, I believe this is not right.

ps: if until now you didn’t read it, here is the paper that started these kind of arguments in my organization: Performing a DBA 1.0 to DBA 2.0 Upgrade

November 18, 2008

Two Days Oracle Index Internals Seminar by Richard Foote

Filed under: Oracle Seminars — H.Tonguç Yılmaz @ 7:34 pm

I have been following Richard’s blog for some time. So it was good to meet him in personal and join his seminar in Istanbul last week. But compared to my previous seminar notes I honestly need to say that I have been lazy this time, sorry.

When you hear the word Internals it is scary for some of us and not so practical for some others. Since I was familiar with the stuff from Richard’s blog the name was not so important for me, but I must mention that this seminar is much more beneficial for a starter to intermediate Oracle DBA profile so the naming can be perceived wrong.

It was long two days, 720 slides of pure Index topics within 10 sessions. But Richard is really a good presenter, with his jokes, up and down waving voice and body language he never let the class to go to a little sleep, even after a Turkish lunch. :)

I also wrote several Index related posts before, it was a good chance to ask Richard the question: Why Indexes? Not Tables, Clusters or Streams etc. His answer was, no specific reason. :) Marketing guys call this kind of a strategy as positioning, I like this very much: Volvo>Security, Lewis>Optimizer, Google>Simplicity, Dyke>RAC, Apple>Flexibility, Millsap>10046 etc. We may say if you have a seat on your mind for Oracle Indexes Richard wants to sit there, I guess.

Within two days Richard discussed on below topics:
– Introduction to Oracle Indexes,
– B-Tree Index Internals,
– Indexes and Statistics,
– Indexes and Constraints,
– Rebuilding Indexes(ironic chapter name for Richard right? :)
– Indexes and the Cost Based Optimizer,
– Indexing Tips, Tricks and Traps,
– Linguistic Indexes and Case-Insensitive Search,
– Other Index Structures like: Reverse Key, Function Based, IOT, Secondary Indexes on IOT, Invisable and Fake Indexes,
– Partitioned Indexes,
– Bitmap and Bitmap-Join Indexes

And here are my highlights from this seminar.

October 22, 2008

INFO: 2 Day Oracle Index Internals Seminar with Richard Foote at Istanbul

Filed under: Oracle Seminars — H.Tonguç Yılmaz @ 8:05 pm

Richard Foote will be in Istanbul for a two days seminar next month. This is an important opportunity, so if you still didn’t you can book here.

Also you can visit here for a general outline of the seminar. As usual I will be publishing my seminar notes, until 13-14 November you can check the previous ones if you like of course. :)

May 16, 2008

Oracle Database 11g Roadshow and Release 2 Sneak Peek

Filed under: Oracle 11g New Features,Oracle Seminars — H.Tonguç Yılmaz @ 4:17 pm

After seeing Pawel’s post I also registered for the event. This three hours briefing was mostly about Oracle Database 11g and provided an opportunity for me to meet with Oracle server technologies vice president Mark Townsend, identify how we can be involved in the 11g Release 2 beta program and outline how we can participate in the marketing of our products and services at Oracle OpenWorld as a partner.

In order to attend you should be in one of below categories;

– Independent Software Vendors
– Systems Integrators
– Consultants, and Pre-sales Consultants
– Resellers

And below are the scheduled next stops;

May 19, 2008: Czech Republic, Prague
May 20, 2008: Poland, Kraków
May 21, 2008: Sweden, Stockholm
May 22, 2008: Spain, Madrid

My impressions is that Pawel was right about his expectations since it was mostly just another colorful Oracle powerpoint day. I tried to ask several questions and took the advantage of speaking to Mark and these were the seconds where we got some technical. Since I believe it can be not appropriate to talk about 11g R2 in details I will skip details of those parts of the event and share my remaining notes as usual;


Oracle Database 11g Roadshow and Release 2 Seminar Notes

For 11g R2 let me briefly mention that other than impressive fusion middleware new features we the database people will be living the new Grid 2.0 era, ASMing everywhere, single RACing to rule them all and edition based redefinition for minimum change impacts on applications ;)

Visit the Oracle PartnerNetwork (OPN) portal

Oracle Database 11g

April 24, 2008

Optimizing Oracle by Desing – my impressions after Jonathan Lewis’s three days seminar

Filed under: Oracle Seminars — H.Tonguç Yılmaz @ 6:29 pm

I have been working on Oracle databases for more than 10 years now. I started with version 7.3 at a local bank and with 8i I moved to the third largest Telco in Europe. If you do value, I also choosed to get certification. I have attended several educations, seminars, conferences, experienced lot of risky projects and weird problems. But my last three days experience was completely something different. Since my native language is not English and this post mostly will not be a technical one, it may get harder for me to express my feelings here.

For years I have been reading Jonathan Lewis‘s articles, books, forum posts and of course after 2007 his blog. In my opinion he is a very special Oracle database expert, the reason why I think this way is not his deep knowledge about the technology but his willingness of sharing all his years of hard work and experiences every time with a very gentle way. Also Jonathan is one of the best teachers I experienced in my whole life. Nowadays it is getting harder to find someone having all this attributes together. IMHO some are very experienced but they do not care about sharing or some “huge mr.ego” experts try to beat people all around forums.

Jonathan’s experiences for all those years produced two important books for the community. But let me talk a little on the seminar book; A4 sized ~400 pages, each page supported with at least three or four paragraphs related to Jonathan’s previous on-site consultancy experiences, which you may not learn from documentation or any google search. These are all nasty things we experience during our daily lives, but Jonathan documented them so well that this course book’s content is like a best of summary of my all time best Oracle books. Even very advanced internals were explained with perfect simple powerpoint slides and supported with example scripts which are all provided in a CD.

Normally after a seminar I share my seminar notes(highlights, topics I liked most lets call), but after this kind of a knowledge storm this time I do not have the hearth for it. Instead I will be sharing the course topics briefly, the new questions I have now after the seminar and some hints you may find useful if you want to attend this seminar in the future.

Yes I have lots of new questions now, normally I love this kind of a feeling, having questions is always good for motivation. But this time I also feel some fear(this may not be the exact word to define, anyway), because Jonathan showed me that even on fundamental topics like redo or undo I still have lots of research to do. This is so normal I guess, because I am limited with the experience that I can get from the databases around me. When operational stuff takes lots of time there is no space for new experiences usually. So for sure I may say in consultancy market you get to learn and develop a lot. I admired Jonathan very much and I felt strongly that in near future I will love to accept its risk and quit to do much more dedicated research on Oracle database.

Course table of contents
Some hints you may like
Lots of new questions I have (if you feel you know some answers please comment :)

Jonathan also uses some extreme examples on purpose from his past experiences, I believe to have and keep the interest all during his long sessions, this way of teaching is something I sympathized. Jonathan shared the SQL scripts to reproduce the outputs for our own databases which he might have tested on 9i or his laptop and shared the results.

I think Oracle community must feel very lucky to have some independent(non-Oracle employee) experts who are being honest, also mentioning the dark sides(bugs, don’t DOs) of this great software, advise to test in details, to be doubtful about so called “guru”s and each new feature for our own comfort. This is unique and precious in todays “marketing” world. For Jonathan’s efforts on developing this course I can not find the right words to explain how thankful I am. I will strongly suggest for the readers of my blog who feel them as an experienced Oracle DBA to attend his seminar at any cost.

ps1: also Graham Woods was in Istanbul, he did a very similar presentation he did on UKOUG last December which I already shared my notes before here.

ps2: UKOUG 2008 call for papers submission deadline for Oracle technology category is getting closer, 9 May 2008. This year I also want to come front, do at least one presentation, I have RMAN, APEX and of course Tuning related topics in my mind. So let’s come front, share our experiences with world’s largest independent Oracle User group community.

ps3: In my 20th month now I am very close to my 200,000th hit(wordpress’s counter). Recently my wordpress dashboard automatically evolved somehow and now I have lots of new stats. As far as I see people once loved event 10046 tracing, flashing back options after 10g and of course my initial best practices posts and no matter what I post after these three seems to be dominant and to stay on top visited list forever I guess. So anyway, thank you for reading and I really hope to have more comments in the future since sometimes I really feel I am alone here :)

March 26, 2008

Graham Wood and Jonathan Lewis will be in İstanbul next month!

Filed under: Oracle Seminars — H.Tonguç Yılmaz @ 3:39 pm

I saw Marco Gralike’s announcement so I also wanted to make a similar one. But this time Graham Wood will visit my company and we planned to be together with Wood for a half day on 21th of April.

Also on 22nd of April Jonathan Lewis will be in İstanbul, this time for his three days seminar. You can find the details of this announcement here. Below are my favorite links for summaries of Lewis’s previous seminars;

https://tonguc.wordpress.com/2007/03/01/another-great-undocumented-oracle-hint/

http://orajourn.blogspot.com/2007/11/day-1-with-jonathan-lewis-execution.html

http://orajourn.blogspot.com/2007/11/day-2-with-jonathan-lewis-writing.html

http://orajourn.blogspot.com/2007/11/day-3-understanding-and-assisting-cbo.html

Next Page »

Blog at WordPress.com.