Difference between count and actual records

[virgo]

People asked for a count of things in a particular file. <br><br>count for rcode=&quot;A&quot; gave 2771 records.&nbsp;&nbsp;The same things in Visual FoxPro version 6 gave 2774 records.<br><br>However, when I retrieved the data from the same file...<br><br>Select * from file where rcode=&quot;A&quot;, I only got 2338 records.<br><br>What caused the difference?<br><br>

 

[dgrewe]

I can understand the difference of a couple of numbers, 2771 and 2774, because of file locking and set delete commands.&nbsp;&nbsp;FPW sometimes does not access records that are locked.&nbsp;&nbsp;&nbsp;A good example of this is to open to FPW table and lock a record.&nbsp;&nbsp;Then open the table again and print a report. The record that is locked will not be on the report.&nbsp;&nbsp;This was fixed in VFP5.&nbsp;&nbsp;Try the test again on a back up copy of the tables so enviroment around the table can be controlled and the test conditions are the same.&nbsp;&nbsp;No one else is using the file and set deleted is off in both FP versions.<br><br>As far as the select statement.&nbsp;&nbsp;My FPW experience is exactly as you discovered.&nbsp;&nbsp;The results vary to much from run to run. I stopped using the select command in FPW and brute force coded what I wanted. I also found out the old brute force method runs faster then the select command.&nbsp;&nbsp;VFP is better at the select statements if the SQL is saved in the DBC.&nbsp;&nbsp;My experience still finds that, even in VFP, that the old brute force method is faster that an off-the-cuff select command. <p>David W. Grewe<br><a href=mailtoave@internationalbid.net>Dave@internationalbid.net</a><br><a href= > </a><br>

 

[JDurbin]

Brute force?&nbsp;&nbsp;like what David?&nbsp;&nbsp;<br>I was unable to duplicate inaccurate results with either RLOCK or DELETE regardless of the SETting of DELETE.&nbsp;&nbsp;But I am a loser&nbsp;&nbsp; <p>John Durbin<br><a href=mailto: > </a><br><a href= > </a><br>ICQ VFP ActiveList #73897253