A good method, test performance under the worst case scenario.

________________________________
From: [email protected] [mailto:[EMAIL PROTECTED] On Behalf Of MDRD
Sent: Tuesday, November 25, 2008 11:38 AM
To: RBASE-L Mailing List
Subject: [RBASE-L] - RE: Which method is best?

Dennis

I will try that but I think I need to set up a slow network to make testing 
easier.
With a fast computer / network it is hard to see the difference in speed.

Thanks for the suggestions
Marc

----- Original Message -----
From: Dennis McGrath<mailto:[EMAIL PROTECTED]>
To: RBASE-L Mailing List<mailto:[email protected]>
Sent: Tuesday, November 25, 2008 11:27 AM
Subject: [RBASE-L] - RE: Which method is best?

I would not do that.

Perhaps you can devise a way to drop and recreate the temp table periodically.

Creating a temp table is faster if you are projecting it from an existing table 
rather than using a create command.

You could create an empty permanent table with the required structure and 
project the temp table from that each time.


________________________________
From: [email protected] [mailto:[EMAIL PROTECTED] On Behalf Of MDRD
Sent: Tuesday, November 25, 2008 10:56 AM
To: RBASE-L Mailing List
Subject: [RBASE-L] - RE: Which method is best?

Dennis

There is only 5-6 rows in the Temp table at a time.
Creating the Temp table each time the form is used seemed too slow. At least on 
older slower computers.

Maybe Packing the Temp table every so often will help?

Thanks
Marc



If you have SCRATCH set to a local drive then Temp tables are stored locally 
and are very fast.
I recommend using indexes (create after populating a table) because this makes 
them even faster and more robust.

Instead of Deleting all rows from a temp table, drop the table and recreate it.
This will prevent the temp file from growing through the day.
Most databases grow in file size when records are added, but do not shrink when 
record are deleted.
This is true also of temporary tables.

I suspect your main slowdown issue is the growing scratch file size.

Dennis McGrath

________________________________
From: [email protected] [mailto:[EMAIL PROTECTED] On Behalf Of MDRD
Sent: Sunday, November 23, 2008 8:45 AM
To: RBASE-L Mailing List
Subject: [RBASE-L] - Which method is best?

Hi

I use to use a permanent table then switched to a Temp table thinking it was a 
better way to go.

Since I can not have indexes with a Temp table I an thinking the permanent 
table might be faster,
especially on a slow network. (Staticdb On)

If I use the Temp table I create it each time the App is started, most offices 
Exit Rbase at lunch and
restart it when they get back.  A few offices said the form that uses this 
table gets slower as the day
goes on.  My thinking it is because of the lack of indexes?

When we start the form I
Delete All Rows in the Table
Insert Rows from the Hist table for the last invoice for that custnum
Edit Using the form Where Custnum = .vcust..make changes if needed.....there is 
3-4 rows in the form
Click a button that Inserts the Rows into the Hist table
Exit the form

Without indexes would the Temp table get bogged down with all the deleted rows 
that linger around
until a Pack or drop and recreate the Temp table?  I tried creating the Temp 
table each time but that
slowed things down.

I thought the Temp table would be better with multiple users and there would be 
less chance of users
deleting the other users rows since the Temp table is local on each computer.

If I use a permanent table I can Pack the Table each time they start the App, 
which is morning and noon
for most offices.

Thanks for any advice
Marc







Reply via email to