Stan, Claudius, your feedback on fc1 ?

[Ivan Smahin]

Hello Ruslan,

Wednesday, December 28, 2005, 12:00:57 AM, you wrote:

RZ> On 12/27/05 9:41 PM, "Claudius Sailer"
RZ> <Claudius at sailer-online.de> wrote:

>> Solution has 2 SQLs which fills an other table from tables and SQL
>> which reads from this table and generates a cursor. And this is
>> faster then having 1 SQL which reads from tables and generates
>> cursor. I would expect that second way should be much faster as
>> solution 1 with V4RB1.10

RZ> You know, this depends.

RZ> So in 1.x you have self did first query and COPY its result into tmp table,
RZ> Then do second query and APPEND its result ?

RZ> Deal is that UNION in v2 follow SQL92, and should do something more complex
RZ> than just APPEND of second result.

RZ> If I not mistake, we need sort both results, to exclude from final result
RZ> duplicates. 

RZ> Am I right, Ivan ?

Yes, it is true for simple UNION, INTERSECT, DIFFERENCE.
But note - if you don't need result without duplicates you can use
UNION ALL and so on...
Here is no sorting/distinct overhead.


-- 
Best regards,
 Ivan                            mailto:IvanSmahin at public.kherson.ua