Re: Tool to check why queries miss results?

From: DA Morgan <damorgan_at_psoug.org>
Date: Wed, 08 Mar 2006 11:36:52 -0800
Message-ID: <1141846598.679119_at_yasure.drizzle.com>


Andreas Koch wrote:
> Sybrand Bakker wrote:
>
>

>>> A) This is the Query i executed
>>> B) This is what results i got (the tool can get that
>>> itself of course)
>>> C) This is a value i would have expected to get, but
>>> didn't.
>>>
>>> I don't look for such a tool because i couldn't debug
>>> such a problem by hand. I look for such a tool because
>>> i'd prefer to debug such a problem in 5 seconds instead
>>> of 5 hours. And if there is no such tool out there,
>>> i'll probably go and write one myself, sooner or later.
>>
>> If the end-user/developer  doesn't supply the 'correct' values in C,
>> what is going to happen?

>
> What are not 'correct' values?
> The values i'm after are coming from some base tables. I'm not
> after "why does select a+b*b+c*d never return 42" or any form of
> solver that could solve all possible problems.
>
> Just one limited problem : Row from one sub-table doesn't cause
> a row in result set.
>
>
> User types values that aren't in the base tables :
> Tool tells him they aren't in the base tables
>
> User types values that are actually returned:
> Tool tells him they are returned
>
> User types values that because of error or wanted filtering aren't
> returned:
> Tool tells him why the aren't returned
>
> I don't understand your problem...
>
> (And yes, this tool isn't suitable as helpdesk replacer. Its just
> a decent data orientated debugger)

Let us cut to the chase. The tool does not exist. In the forseeable future the tool will not exist. The proximate issue is that your people need training and you should endeavour to fix the problem rather than look for some theoretical solution.

[Quoted] So continue to beat this to death if you wish ... but it doesn't exist. And I talked to an associate that builds AI systems and he said that if your budget was in excess of $200K US he would entertain evaluating whether it could be built. Not build it ... just explore the possibility.

-- 
Daniel A. Morgan
http://www.psoug.org
damorgan_at_x.washington.edu
(replace x with u to respond)
Received on Wed Mar 08 2006 - 20:36:52 CET

Original text of this message