Re: Multiple datasets in one dataset

On Thu, 11 Oct 2007 00:29:30 -0700, roman.morokutti_at_googlemail.com wrote:

>Hi,
>
>> For a quick solution, you could use a UNION ALL, and a WHERE NOT
>> EXISTS on each one.
>>
>
>thank you for your help. Could you give me an example of how such
>a statement could look?

select * from foo where wp = '418' and curr_ms = 28144336 and value_group = 11
UNION ALL
select * from foo where wp = '418' and and curr_ms = 28144336 and value_group = 21
UNION ALL
select * from foo where wp = '418' and curr_ms = 28144336 and c.value_group = 22

Of course a standard IN() would work as well.

>
>
>
>
>> I fail to see how do you know that: a.value_group = 11 and
>> b.value_group = 21 and c.value_group = 22. And if you do know there
>> separate values, why not just write separate queries?
>>
>
>This is as I stated above that it is not clear how many lines to
>expect.
>The example shows a query where I definitely know the current count
>of lines. The lines indicate a process flow or think of a pipeline
>where
>a piece comes in and goes through the stages. The initial stage
>creates
>the unique time stamp and sets the value_group to 11. The next stage
>takes the same time stamp and sets value_group to 21 and so on till
>the last stage and this has the value_group id 25.
>
>But if the processing in one stage before the last may fails I have
>occasionally not the full count of entries. And this is exactly the
>problem.
>To select all them regardless of on what stage they may have been
>failed.
>
>Regards
>Roman
>

I still do not understand. What defines if another record should be searched? What denotes failure?

B. Received on Thu Oct 11 2007 - 08:38:40 CDT