What is taking all the time? Are you meaning that to run a query against
your large sets which would produce 100 rows of data is taking 14 hours or
to get the 10M rows down to the PC is taking 14 hours? Have the query run on
the HP and only send the response. Limit the responses to a few records.
Introduce keys and indexes where it would be the most good, and finally look
into wharehousing. Pre-groom the data from the large sets and post the
responses to a smaller set and only allow the users to access the groomed
data.

James

> -----Original Message-----
> From: HP-3000 Systems Discussion [mailto:[log in to unmask]]On
> Behalf Of [log in to unmask]
> Sent: Tuesday, May 09, 2000 2:16 AM
> To: [log in to unmask]
> Subject: Once more: Connecting HP to the PC-World, DataExchanges and
> ODBC-Interfaces
>
>
> Hi to all!
> Often in this group there is a discussion what is the best
> solution to do a
> connection between PCs and the HPs, doing data-exchange (from
> TurboImage-Databases) and which ODBC-Interface works the best.
>
> We tried several Interfaces and possibilities but we were not
> able to find a
> usefull solution. Our databases are very big (up to 50 Million entries,
> cross-connected and... and... and...). To offer only the customer
> dataset we
> use takes about 14 hrs to proceed using the HP-ODBC-Interface...
> The Order-Headers take nearly the same and the Sub-Headers have
> taken about
> 24 hrs to offer by ODBC... (10 Million entries...).
>
> These times are unacceptable, but we can't decrease them...
> We also tried the Minisoft-ODBC-Interface but with the nearly
> same results.
>
> Our system is a HP979 and we want to be able to access the datas on the
> PC/Windows-Side (Reporting, Analyzing,...).
>
> Has anybody a hint or solution what could be the reason or what
> else we can
> try?
>
> Thanks a lot!
>
> Holger Schrörs
> QVC-Germany / IS&T - MIS
>