Database Concepts & Design | Forum profile

To answer your question properly would require about 20 hours of work. I charge $125 an hour for my time, and my guess is you don't have the $2,500 to cover my effort. Five minute questions I answer for free, beyond that you pay.

Hi, Laurent, thanks for the reply. So what you are basically saying is that there is no nautural key in the customer table. But what about a userId that the user enters. Is that a natural key or not? We must have things that distinguish one customer from the other, that makes them unique. But maybe it is as Mark said earlier. It's up to the user or constratints in the application that will sort out things like customers being named same...

Nearly every database server offers some feature like what you've described. Microsoft also offers clustering, mirroring, and merge replication. Oracle offers RAC. DB2 offers clustering. Sybase offers several solutions, based on which database engine you're using. My experiences with MySQL clustering haven't been very good. I'd suggest that you at least review your other options before you make a commitment. -PatP

Quote: Originally Posted by verellen Hey all, I have the following question: When designing a database, qua performance, which is the best senario: Having a table with lots of columns, so you do not have to use many joins. Or storing your data in seperated tables, which is more clarifying and flexible, but when you need to gather the data, you'll have to use more joins. When does having a lot of joins have affect on the performance?...

Not a natural key ? Well, if you want, but I am very uneasy with this vocabulary. Because, this concept, as many others, is formally defined on a finite set of things and properties you've defined before, so to speak, the abstract picture of some reality you want to manage. As all abstract models, our softwares don't reflect exactly reality and all its features. We have to simplify things to manage them. How much ? There's no...

Nobody expects perfection. However I do think it's reasonable to ask database software vendors to show some improvement in basic features after more than 30 years! Especially when you consider the outstanding innovations in other fields of IT during that same period. Commercial DBMS technology has stagnated and shown very little improvement in core features during the last three decades. Part of the reason is of course the cosy vested...

By googling, the name Commit ordering is almost equally used for Commitment ordering .

ok so no difference for 0,1 1,1 0,n 1,n

Robbery. I will do your homework for you for $120 per hour. Spell-checking would be extra, though.

Ok, thanks for you reply! It will be kind if you can maybe answer my second question: What about working offline, If I would like to work offline with some SQL server, and when i get online, my application will execute all INSERT Queries. Do you know about any API that do that? virtual server on the client that can do that? or any buffer that can help? Thanks in advance!