Re: Foreign key in Oracle Sql

TML wrote:

> Pardon my jumping into the fray, here, but:
>
> DA Morgan wrote:
>

>>So what is it you think those of us that work in Oracle do? Do you
>>think we are overly burdened by the issue you bring up? Do you
>>really think I have even once written a handler for empty strings
>>in Oracle? Really? Not even once have I had to. Oracle provides
>>the tools to do it and have for years. Look up the built-in
>>functions NVL and NVL2. So much for your "big issue."

No. But you can assume that when I do I get paid for doing so. And you can assume that this is how my mortgage gets paid, how gasoline gets into the fuel tank of my car, how my family is fed, and how I save money for the future. In short ... someone else's lack of being able to read a tech doc is what puts food on the table.

  when their package (or what-have-you) didn't understand and
> account for this "nit"? Sure, I don't have to write NVL, it's already
> there. But I very well MIGHT have to dig through hundreds of thousands
> of lines of code to find the place where some developer forgot to call
> it...or just plain didn't know to call it, as I'm more surprised to
> find a developer who HAS read documentation than I am to find one who
> hasn't. (All praise the CS/IT Diploma mills.)

And why is this a bad thing? Would you prefer unemployement or flipping burgers?

> I know of companies that spent lots and lots of money to clean up
> front-end code that broke because of this neat little trick.

And from your standpoint as a developer why is this a bad thing? From management's point-of-view it may well be but why is that your problem? Let them learn from their mistakes and next time hire you instead.

  Regardless
> of the fact that Oracle provides me a way to deal with this, it's still
> a burden on the DBA that doesn't make any sense to me.

Unless you are a charity or a slave why the attitude toward the very thing that justifies your paycheck?

>>>that is, a compatability setting.
>>
>>Works in end-user code. Doesn't work in kernel code.

Take a look at the bin directory. Note that nice little ~45MB file named oracle.exe (assuming Windows and 10g). It queries the data dictionary all of the time. It uses dual. It assumes a particular behavior for NULL. So does every other piece of executable code Oracle has written. Would you add a compatibility switch to every single one of those statements?

Now go to RDBMS/ADMIN. Grep or search for NVL. I get 60+ files and that is not counting all of the .PLB files that contain wrapped code. And we haven't even begun to discuss code in JDeveloper, Designer, Forms, Reports, Colab Suite, App Server, Oracle Financials, Oracle CRM, Oracle HR, Rdb, ODBC drivers, and myriad other sources. You are talking about a monstrous project.

>>And 100,000 existing DBAs and developers would hate it
>>because they have had it a specific way for 20 years.

For everyone that comes to Oracle and doesn't like changing from what they learned with DB2, Informix, Sybase, or SQL Server you can very safely assume there are at a minimum 10 DBAs and developers that began with Oracle and would like you to stuff a sock in it.

I ran an informal poll at a recent user group meeting. Not one person thought it a good idea.

  It would
> certainly increase the demand for Oracle developers and DBAs, making
> the market more competitive.

Surely you jest. It would only hurt Oracle developers and DBAS. It would open a door for additional competition from people who would only whine about not having autonumbering.

..it would give the people who write Oracle
> support tools, and apps that rest on top of Oracle, another chance to
> ask their customers to pay for the product(s).

Do you live in a cave? Do you have any idea how many third-party companies write tools that run on Oracle? Far more than on SQL Server, Sybase, and Informix combined.

> It seems to me the only real "losers" would be the companies who employ
> Oracle solutions, and how many of the people who face that bottom line
> are going to understand what it is, exactly, that they're getting when
> they put more money into "upgrading Oracle?" My gut feeling is "not
> very many", but that's purely based on anecdotal evidence.

My feeling is that you haven't a clue what Oracle is. You still think it is a relational database ... some form of MS Access or Excel on steroids.

If I told you it was the most heavily instrumented database ever developed likely you would neither understand what I was talking about or why that is of value.

> You can certainly count me as one of the DBAs/developers who wouldn't
> mind seeing this, and lots of other Oracle oddities, go away. :)

Actually my sense is that your opinion doesn't count at all as it is no such a miniscule minority of opinions on the subject. And one that in my mind marks you as someone that has either never worked with Oracle or who knows the product at such a superficial level as to be laughable.

But I'd love to have you prove me wrong so here's a basic question. How would you use a function based index to enhance the retrieval performance of records in a column where the only values are "Y" and "N" and why would a B*Tree, bitmap, reverse, descending, domain, cluster, and/or compressed index not work as well?

-- 
Daniel A. Morgan
University of Washington
damorgan_at_x.washington.edu
(replace 'x' with 'u' to respond)


----== Posted via Newsfeeds.Com - Unlimited-Uncensored-Secure Usenet News==----
http://www.newsfeeds.com The #1 Newsgroup Service in the World! >100,000 Newsgroups
---= East/West-Coast Server Farms - Total Privacy via Encryption =---