Subject | Re: [IBO] Poor Performance of Large Datasets |
---|---|
Author | aldushuxley_29 |
Post date | 2003-10-30T21:43:55Z |
Good Morning Artur,
I am sorry for not responding sooner, I had some problems with my ISP
yesterday. The area that I have noticed which is slower is where I
filter on a dataset. Specifically this is where I am performing a
filter on my dataset. I do a containing search on a varchar(30)
field and it is very slow.
Comparing my converted application (this is a job costing / process
costing system so it has many records) back against the BDE
everything takes about the same time, except this search routine
which takes much longer (10 seconds instead of 2 seconds). Would the
fact that I have not noticed a performance gain running across the
network indicate that I have BDE settings interferring with IBO? If
so is there anything that people have encountered before that they
could identify as being redundant or negative in its effects on IBO.
Your thougts are most welcome.
Cheers
Aldus
I am sorry for not responding sooner, I had some problems with my ISP
yesterday. The area that I have noticed which is slower is where I
filter on a dataset. Specifically this is where I am performing a
filter on my dataset. I do a containing search on a varchar(30)
field and it is very slow.
Comparing my converted application (this is a job costing / process
costing system so it has many records) back against the BDE
everything takes about the same time, except this search routine
which takes much longer (10 seconds instead of 2 seconds). Would the
fact that I have not noticed a performance gain running across the
network indicate that I have BDE settings interferring with IBO? If
so is there anything that people have encountered before that they
could identify as being redundant or negative in its effects on IBO.
Your thougts are most welcome.
Cheers
Aldus
--- In IBObjects@yahoogroups.com, "Artur Anjos" <artur@a...> wrote:
> Hi,
>
> > In this case
> > I need to use the Dataset descendants as rewriting the
application is
> > not an option.
>
> I use TDataset descendants as well.
>
> > In some screens I have a fairly large amount of data
> > (5000 to 15000 records, and yes I know it is against Client/Server
> > design principles, but this is a legacy system),
> Well, I don't have so many data in any screen...
>
> > and at present it is
> > slower for these large datasets than the BDE.
> Keylinks will be your friend here. Good index as well.
>
> Share with us where do you have slower results. These should always
be small
> changes in your code to adapt.
>
> Artur