Hi Jeff ---

Without data buffering (but Jose said he was buffered, hence my response) I have also run into the problem. I think it has to do with Windows read/write caches and internal buffering. I have never been able to eliminate the problem but I did reduce it by disabling the write cache in Win95 one time.

I don't thnk the problem can be eliminated. I suspect that some call in Windows reports that something is physically written when it's not...so the report is milli- or microseconds off.

I do what you do and just re-open the durn thing. But that's not an option (at least a practical one) if it's a multi-user table and there is a lot of traffic. In those circumstances, I abandon DBFs altogether and go to physical n-tier with SQL Server or Oracle.


>OK John splain this one to me.
>The reason I responded the way I did was I ran into the same problem. NO BUFFERING. Direct table access. Table was on a server, select 0,use table,
>this was a table that held one field and one record, one field and record only , always going to be one field and one record. Its a datetime field of the last time a process was run. Rlock did not do it for me. The only way I could get the new value was to
>a. close the dbf and open again.
>b. add another record that was blank and toggle between the two.
>
>I assumed it was because of buffering on the network and VFP's loading into memory. I guess I could have tryed a flush but did not think about it at the time. What else could I done?
>
>Jeff
>
>>Hi Jim ---
>>
>>I don't think he wants to "GO" anywhere. If it's row buffered. What I'd like to know is if this table will stay one record.
>>
>>>
>>>If it has only 1 record then
>>>
>>>GO 1
>>>
>>>should work just as well too.
>>>
>>>Cheers,
>>>
>>>Jim N
>>>
>>>>I Got it
>>>>
>>>>I forgot to say the table has only one record
>>>>
>>>>I did this:
>>>>
>>>>Select Table
>>>>SKIP && Go EOF
>>>>SKIP -1 && Return to the record
>>>>
>>>>An it works great
>>>>
>>>>Thanks to everybody