Query or Select (AREV Specific)
At 26 JUN 2002 06:46:04PM Patrick Caufield wrote:
Arev 2.03
NT 4.00 server
workstation win 2000 pro
This is probably a dumb question.
When using window query or a select statement in a program the system will not return a specific record(s). In a file select records having a contract no of 2353.22611 (this is not the record key) additionally there are records that have a contract no of 2353.2261. The records returned include both contract numbers. The field data type is float left justified.
Is this normal. If not how do you select the only contract no. 2353.22611?
Thank You in advance for any help.
Pat@escrowinc.com
At 26 JUN 2002 10:06PM Curt Putnam wrote:
Is the field indexed? If so, you might want to check the field length and ensure that it is about 5 or so characters longer than the longest possible value. If contract number is not operated on numerically (even if it is) consider making it a character string.
At 27 JUN 2002 11:31AM Patrick Caufield wrote:
Curt
Thanks for the help
Additional information
We collect payments on owner financed notes therefore;
1. The contract no is indexed in most files.
2. Currently the max length is 10 with a field length of 13.
3. The contract no is the primary reference no in the system.
4. Almost all the files contain the contract no.
5. Only one uses the contract no as its key, that file holds all the data required to process payments for each contract.
I guess this would mean alot of work or live with making sure that this sequencing of the contract numbers is not allowed.
Pat
At 27 JUN 2002 09:48PM Curt Putnam wrote:
In order to speed up indexing, at some point Arev added a requirement to make your field length in the DICT actually relate to the data. You will have to reset the field length in the DICT to max length plus a few. I suspect you will also have to delete & recreate each index. That should solve the select problem.
At 28 JUN 2002 07:07AM [url=http://www.sprezzatura.com]The Sprezzatura Group[/url] wrote:
If you changed the key length after creating the index, you must remove then add back the index.
This field, is it a number or a string? By number, I mean are the values always numbers, like in your example. Are there any format conversions (ICONV/OCONV) on the dictionary.
If it's numeric, you might want to consider changing the justification to R. The sort collation will change if you do this.
You might also want to consider using the num_to_char conversions on this field, which will treat the code as a string.
I think you're losing some information on precision issues.
For more on precision you should read Revmedia Vol3, Issue 7, pages 8-11.
World Leaders in all things RevSoft
At 28 JUN 2002 08:33PM Patrick Caufield wrote:
THANK FOR THE HELP I WILL TRY THESE IDEAS AS SOON AS POSSIBLE. LUCKILY THE PROBLEM IS CRITICAL.