DBAsupport.com Forums - Powered by vBulletin
Results 1 to 7 of 7

Thread: The comparison of SQL Server 2000 with Oracle 9i

  1. #1
    Join Date
    Apr 2003
    Posts
    1

    The comparison of SQL Server 2000 with Oracle 9i

    has anyone seen a technical document that compares SQL Server 2000 with Oracle 9i

  2. #2
    Join Date
    Sep 2000
    Location
    Chennai, India
    Posts
    865
    Check these out...

    Oracle beats SQL Server 2000 in Stored Procedure Programming...
    http://www.devx.com/dbzone/Article/11363

    Database study puts Oracle on top ...
    http://www.itworld.com/AppDev/119/020820databasestudy/

    Clash of the Titans - Oracle on top...
    http://www.pcmag.com/article2/0,4149,7279,00.asp

    HTH.

  3. #3
    Join Date
    Nov 2000
    Location
    Baltimore, MD USA
    Posts
    1,339
    http://www.devx.com/dbzone/Article/11363 <- Wow! That was as shallow an investigation as I've ever seen.

    Mind you, there's not much out there, so that he addressed the subject at all is commendable, I suppose, but I'm just not comfortable with his conclusions. There are soooooo many variables is the problem. Now, everything he said was true, AFAIK, and his conclusions, had he hedged them a little, were supportable. T-SQL is a weakness in SQLServer and Oracle's packages and error-handling are, indeed, vastly superior, *but*. He glossed over the learning curve issue too quickly, for one. T-SQL is wholly coherent and clean. PL/SQL is a mish-mash of stupid crap that's been thrown together over the past, what, 20-30 years? Just as a minor digression here, one of the main complaints against Microsoft is also one of its main strengths - they have no problem throwing away the concept of backwards-compatability. Every couple versions, they completely re-write major sections of the codebase. This allows them to continually produce a self-coherent product. PL/SQL, and Oracle as a whole, is more in need of a complete overhaul than my rust-bucket of a 30-year old car was back in college. There are, for example, at least 10 different ways of executing a SQL SELECT statement in PL/SQL:

    1 - in-line INTO
    2 - Cursor FOR Loop
    3 - Cursor linear fetch loop
    4 - NDS linear fetch with weak refcur
    5 - NDS linear fetch with strong ref-cur
    6 - DBMS_SQL linear fetch
    7 - in-line bulk fetch
    8 - Cursor array fetch loop
    9 - NDS array fetch with strong ref-cur
    10 - DBMS_SQL array fetch

    Oh, but note that you can't do an NDS array fetch with a weak ref-cur. This is how all of Oracle and PL/SQL is: Way too many options with far too many of the combinations not working. You'll find that almost every command in PL/SQL or SQL has a set of situations where it won't work. Why? Because every new piece of functionality is 'added' to the outside and has to be hooked to 500 existing pieces of functionality. How does one solve this problem? Every so often, you have to step back and re-design the beast from the ground-up to handle all the desired functionality in a self-coherent and intuitive fashion. Microsoft does it all the time, which is why their products are orders of magnitude easier to learn than Oracle's. Oracle has yet to do it once.

    So yeah, you can say that PL/SQL is more powerful than T-SQL, but the price of that power is something that is far too complex and archaic. I'm glad the author thinks that finding a 'good book' makes up for that difference. I do not.

    Don't get me wrong - I like the power that Oracle provides. I've paid my dues to learn it as well as I have, however, and I keep paying them every day. It would just be nice if Oracle would take a course or something on how to not only make a powerful product, but a user-friendly one as well.

    So yeah, PL/SQL has packages and error-handling, which are very nice indeed, but I simply don't think that looking at 2/600 variables is a sufficient basis for then uniquivocally declaring that PL/SQL is the best environment. He should have hedged that conclusion a little more, IMHO.

    - Chris

    (damn, but I'm picky )
    Christopher R. Long
    ChrisRLong@HotMail.Com
    But that's just my opinion. I could be wrong

  4. #4
    Join Date
    Nov 2002
    Location
    Geneva Switzerland
    Posts
    3,142
    Chris, did you note the reviewer's profile?
    "Joe Lax has spent more than 10 years working in various database environments and has been a practicing DBA on all versions of SQL Server from version 4.2 onward. He also is a MCSE and an MCT who recently has started to learn Oracle, which affords him no end of fun."

    Perhaps the real truth about T-SQL is even worse that the real truth about PL/SQL? Or is this a case of "the grass is always greener"?
    "The power of instruction is seldom of much efficacy except in those happy dispositions where it is almost superfluous" - Gibbon, quoted by R.P.Feynman

  5. #5
    Join Date
    Aug 2002
    Location
    Colorado Springs
    Posts
    5,253
    Originally posted by chrisrlong
    Just as a minor digression here, one of the main complaints against Microsoft is also one of its main strengths - they have no problem throwing away the concept of backwards-compatability. Every couple versions, they completely re-write major sections of the codebase. This allows them to continually produce a self-coherent product.
    Haven't worked with SQL Server, but am I right in inferring that Sql Server applications may need re-write when the code base gets changed?
    David Aldridge,
    "The Oracle Sponge"

    Senior Manager, Business Intelligence Development
    XM Satellite Radio
    Washington, DC

    Oracle ACE

  6. #6
    Join Date
    Nov 2000
    Location
    Baltimore, MD USA
    Posts
    1,339
    Yes - they have made major changes to the code in releases 6.0 and 2000.

    At these major releases, there were many non-trivial changes that had to be made to code in order to upgrade. Now, if a database is properly insulated with an abstraction layer, this is not as painful as it sounds. Microsoft has done this in every product they've made. A program written in Visual Basic 3.0 cannot possibly run in the latest version - the language is significantly different.

    SQLServer started as a joint venture with Sybase, back in version 4.2.1. The entire engine was re-written in 6.0. Major optimizations were implemented in 6.5, and on and on. So yes, there are often major changes that are involved in upgrading to a new Microsoft release. As I said, this is a major complaint against Microsoft. However, I also believe it to be one of their major strengths as well. Yes, making changes for an upgrade can be painful.

    But the other extreme is to forever do your utmost to maintain backwards-compatability and keep slapping on new patches and functionality to the same old code. Oracle is an extreme example of that approach. It has so much Duct tape and so many rubber bands holding it together that it seems a miracle it works at all sometimes. Any time I venture into new territory I suffer major frustrations. You can now do SELECT-clause sub-SELECTs, but we can't give you the plan. You can now do autonomous transactions, but not in a distributed transaction. You can do array fetching, but not on a weak ref cursor. You can treat an array as a table, but not in a distributed transaction. And on and on and on.

    Oracle is massively in need af a paradigm-shift. This command-line, circa 1950 feel of everything is simply pathetic in this day and age. UTL_FILE, for example, is the most idiotic piece of garabage I've seen in a long time. It looks and acts nothing like any of the other packages. It feels as if it was written by some high school drop-out 30 years ago and we've been stuck with it ever since. Heaven forbid they change something in Oracle, people would have a fit.

    But then again, I doubt Oracle would make a good paradigm even if it did start again. Oracle's persistent weakness is its inability to make any product usable. It's applications are pathetic. Compare SQLServer's interface to Enterprise Manager sometime. Their apps, their languages, their packages and their utilities all suffer from massive usability issues.

    Now, I have chosen the Oracle camp. I like the power of the database. I switched from SQLServer - my previous expertise - several years ago. I complain like this because I care. I also worry that if Oracle doesn't address the usability issue directly, and not with the half-steps they have thus far, the stock-swap ratio when Microsoft buys them out will not be good

    - Chris
    Christopher R. Long
    ChrisRLong@HotMail.Com
    But that's just my opinion. I could be wrong

  7. #7
    Join Date
    Jan 2001
    Posts
    3,134
    Well don't hold back Chris, tell us how you really feel.




    Very good points by the way.

    MH
    I remember when this place was cool.

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •  


Click Here to Expand Forum to Full Width