This is a discussion on READ_UNCOMMITTED problem with SQL 2000 and i-net Opta 2000 JDBC within the SQL Server forums, part of the Microsoft SQL Server category; --> Dave Hau wrote: > Daniel, I think what rk is talking about is whether Oracle ever allows you > ...
| |||||||
| FAQ | Members List | Calendar | Search | Today's Posts | Mark Forums Read |
| ||||
| Dave Hau wrote: > Daniel, I think what rk is talking about is whether Oracle ever allows you > to read a block without *checking* its SCN, not just reading the SCN. Of > course, you're right - when you read a block, you read the SCN as well. But > he's talking about the extra step of checking the SCN against the SCN at the > beginning of the statement execution (for statement level read consistency) > or the SCN at the beginning of the transaction (for transaction level read > consistency) to determine if you need to roll back the block. > > Regards, > Dave The SCN is only checked in transactions where the data could have been modified between the initiation of the transaction and when the cursor gets to the rows. One example of the SCN being ignored is in read-only tablespaces. Another would be when rows are intentionally locked with SELECT FOR UPDATE. -- Daniel Morgan http://www.outreach.washington.edu/e...ad/oad_crs.asp http://www.outreach.washington.edu/e...oa/aoa_crs.asp damorgan@x.washington.edu (replace 'x' with a 'u' to reply) |
| |||
| "Daniel Morgan" <damorgan@x.washington.edu> wrote > The SCN is only checked in transactions where the data could have been > modified between the initiation of the transaction and when the cursor > gets to the rows. One example of the SCN being ignored is in read-only > tablespaces. Do datas get moved to read-only tablespaces magically. If I have a workflow table where rows only get inserted/updated for a short period of time, and then remain static, it makes sense to use READ UNCOMMITTED when running reports for anything other than the current activities. I don't know how easy it is in oracle to move yesterday's data in a read-only tablespace. >This is an impossible question to answer. Because you can not ever read >a block (in SQL Server verbiage page) without reading the SCN as it is >stored in the block. What about comparing the SCN in data block with the SCN when the query was initiated. As I see, unless it is a read only tablespace or SELECT FOR UPDATE, it always does this SCN comparison. It is this SCN comparison which I think is totally unnecessary when dealing with static data as in the above example I gave. rk- |
| |||
| rkusenet wrote: > "Daniel Morgan" <damorgan@x.washington.edu> wrote > >>The SCN is only checked in transactions where the data could have been >>modified between the initiation of the transaction and when the cursor >>gets to the rows. One example of the SCN being ignored is in read-only >>tablespaces. > > > Do datas get moved to read-only tablespaces magically. Can you ask a question without being belitteling, demeaning and/or insulting? I'm serious: Because the answer to this questions is blazingly obvious. My contributions are an attempt to educate not some testosterone laden internet contest. I invite you to reask your question with more appropriate verbiage if you actually want an answer. If not ... you win, I lose, you're the man. -- Daniel Morgan http://www.outreach.washington.edu/e...ad/oad_crs.asp http://www.outreach.washington.edu/e...oa/aoa_crs.asp damorgan@x.washington.edu (replace 'x' with a 'u' to reply) |
| |||
| "Daniel Morgan" <damorgan@x.washington.edu> wrote > Can you ask a question without being belitteling, demeaning and/or > insulting? I'm serious: Because the answer to this questions is > blazingly obvious. > My contributions are an attempt to educate not some testosterone laden > internet contest. I invite you to reask your question with more > appropriate verbiage if you actually want an answer. I admit that the tone of my post was derisive and it was intentional. Bcos I find you highly opinionated and the way u insult non oracle products is offending. Be prepared to get offended if u offend others. You find faults with all products and jump in to defend Oracle the moment it is attacked, making ur bias too obvious. For e.g. you mentioned recently that SQL Server is so bad that even a non tuned Oracle/Informix/DB2 can give performance upto 9x times. Bah. I agree that SQLServer is not yet in the class of Informix, but to say that it is way behind them is plain stupid. Do u think some of the heavily used sites like BarnesAndNoble, Dell,Expedia are stupid to use SQLServer. or Verizon Communications, Lucent etc. check this: http://www.wintercorp.com/vldb/2003_...TenWinners.asp You can disregard my question re: oracle read-only tablespace. At this time I have no desire to learn about Oracle. I sincerely offer my apology if I offended you and I hope that you will also change your attitude towards non oracle products. cheers. rk- |
| ||||
| rkusenet wrote: > "Daniel Morgan" <damorgan@x.washington.edu> wrote > > >>Can you ask a question without being belitteling, demeaning and/or >>insulting? I'm serious: Because the answer to this questions is >>blazingly obvious. >>My contributions are an attempt to educate not some testosterone laden >>internet contest. I invite you to reask your question with more >>appropriate verbiage if you actually want an answer. > > > > I admit that the tone of my post was derisive and it was intentional. > Bcos I find you highly opinionated and the way u insult non oracle > products is offending. Be prepared to get offended if u offend others. > > You find faults with all products and jump in to defend Oracle the moment > it is attacked, making ur bias too obvious. For e.g. you mentioned recently > that SQL Server is so bad that even a non tuned Oracle/Informix/DB2 can give > performance upto 9x times. Bah. I agree that SQLServer is not yet in the > class of Informix, but to say that it is way behind them is plain stupid. > Do u think some of the heavily used sites like BarnesAndNoble, Dell,Expedia > are stupid to use SQLServer. or Verizon Communications, Lucent etc. > > check this: http://www.wintercorp.com/vldb/2003_...TenWinners.asp > > You can disregard my question re: oracle read-only tablespace. At this time I > have no desire to learn about Oracle. > > I sincerely offer my apology if I offended you and I hope that you will > also change your attitude towards non oracle products. > > cheers. > > rk- I'm not in marketing and make no more money from Oracle than I do from other RDBMS products. That I teach one at the U says more about what the students want than anything else. And I have repeatedly pointed out that different RDBMS products are different. I don't rush to defend or attack any of them because I have no vested interest in any one of them. If you insist on putting a value judgement on differences ... good vs evil, white hat vs black hat, cowboys vs indians, that is your value judgement ... not mine. And I'm not at all offended. I just don't choose to play games that consist primarily of seeing who can have the most testosterone shooting from their eyeballs. As an educator I am surprised you have no interest in learning Oracle. I make it a point to keep up my skills in SQL Server, DB2, and Informix. Used to also keep current in Sybase and Teradata. Much of my consulting work comes from helping people that don't know multiple products perform migrations or make a product work in heterogenous environments. Guess I should thank you for keeping my vacation plans on-track. -- Daniel Morgan http://www.outreach.washington.edu/e...ad/oad_crs.asp http://www.outreach.washington.edu/e...oa/aoa_crs.asp damorgan@x.washington.edu (replace 'x' with a 'u' to reply) |