Very Computer

Hi Paradox-Users,

I have a Speed-Problem with Paradox 4.0 and WinNT4.
Up to now, nobody could help me, to speed up my Applications
and Tables (only by changing NT to Novell 4.xx, but this is not
the way I want to go, because there are many other Apps especially
for NT).
I think, maybe its helpful to reduce the size of my tables (the largest
one is about 108 MB). But I dont know how to do.
Note:    It is neccesary to keep the old data (from 1993 - 1996) for
            further use, to make some statistics and so on. So i need a
            very simple method to merge the old data into my new tables.
            (there is Paradox 5.0 available, for more comfortable working)

Are there any experts, who nows something about
Many thanks

or put them into the newsgroup comp.databases.paradox

Quote:>Hi Paradox-Users,

>I have a Speed-Problem with Paradox 4.0 and WinNT4.
>Up to now, nobody could help me, to speed up my Applications
>and Tables (only by changing NT to Novell 4.xx, but this is not
>the way I want to go, because there are many other Apps especially
>for NT).
>I think, maybe its helpful to reduce the size of my tables (the largest
>one is about 108 MB). But I dont know how to do.
>Note:    It is neccesary to keep the old data (from 1993 - 1996) for
>            further use, to make some statistics and so on. So i need a
>            very simple method to merge the old data into my new tables.
>            (there is Paradox 5.0 available, for more comfortable working)

Questions:
1) have your tables been Normalised? You talk about 1 table and did
not mention record counts.

2) Are your field sizes optimum for the data comtained?  Using Short
(+/- 32,767) numbers can save 6 bytes per record.

3) moving "History" data to another table and only combining the
Active data ANSWER with the history Data ANSWER for reporting .

Steve U

        _
       | \                           Steve Urbach
       |  )erek

  / / /                              http://www.greenduck.com

If you need to continue using the older data, the simplest path is to break
the data into multiple tables, and modify the code so that it hits the
appropriate table(s)

100 meg of data isn't necessarily "bad".. it's quite possible that the way you
use the tables can be optimized.. tighter and/or multiple queries, better
indexes, etc..

--
Steve Green
Corel CTech - Paradox
Diamond Software Group, Inc.
Waldorf, Maryland  USA

http://www.diamondsg.com    <- please note our new web site

Howdy:

You might also look for any other programs that are open on the
desktop - a virus checker set to look at every file prior to allowing
the file to be opened would drive a Gray to it's knees.

HTH.

Bob

It's worth pointing out, for example, that many queries only use indexes
that are "maintained" AND "case sensitive."

Sometimes you are tempted to write one very big complicated query to get
the results you want and find out that Paradox is dog-slow in executing
it.  But if you create a sequence of small queries to do the same thing,
you can get excellent performance.

That is how the report-sequencer we sell as "ReportRabbit" came to be,
and we use the concept all the time.  Our most extreme application had
over 5,000 rows in its sequence table.

On Wed, 13 Jan 1999 06:25:17 -0700, Sundial Services

I will second this theory. Long ago I found that a number of Simple
queries ran much faster than a "Perfectly Valid" complex Query.

In many cases, I do my own "Outer Joins" as they tend to run up to 10
times faster than letting Paradox do them.

Steve U

        _
       | \                           Steve Urbach
       |  )erek

  / / /                              http://www.greenduck.com