Re: Navigation question

On Feb 20, 12:45 pm, "JOG" <j..._at_cs.nott.ac.uk> wrote:
> On 20 Feb, 16:19, "dawn" <dawnwolth..._at_gmail.com> wrote:
<snip>
> > RM isn't navigated (but, at the risk of you wanting to hit me,
> > databases often are). That is why I'm asking such questions. There
> > seems to be a mindset against any form of navigation and I am trying
> > to figure out just what precisely is problematic with navigation in
> > either of the two ways I have indicated. 1) programmer using a value
> > from a resultset to seed another query or 2) DBMS having the
> > specifications to be able to understand the logical navigation.
>
> > Thanks for your patience with me. --dawn
>
> Argghghgh. Now I am become Sisyphus, the pusher of stones.

Rats, sorry to frustrate you such, JOG. I am doing what mAsterdam asked and trying to figure out if I missed or misunderstood any "answers." I am still struggling, as you can tell, with just what precisely is problematic about "database navigation."

I'll give it another shot and hope you will too. Perhaps I should ask if you are willing to tell me what form of "navigation" (sans iteration) is problematic. I have given two examples that I think you think are problematic, but where I do not grok the problem.

1. Output from one SQL statement serves as input to another: Developer does something like select customerid, customername from Customer where tin='xyz'; somewhere along the line, assigns the value of customerid to a variable, such as thisCustomerid and later (without any user input) issues select orderid from Orders where customerid=thisCustomerid;
2. DBMS permits link specification Developer specifies a link, such as Order to Customer on customerid Then issues statement like the pseudocode: select orderid, orderPrice, Customer.customerName from Order;

The DBMS has the spec so it "knows how" to get "from Order to Customer"

Is either or both of these an anti-pattern from your perspective and, if so, why precisely?
Thanks --dawn Received on Wed Feb 21 2007 - 04:31:49 CET