>
>I use tableupdate. No flush. I am on an NT 4 SP3 server.
>
>Below is my response to Peter. It enters the data but does not update the CDX.
>
>I was able to do some more research on this after I sent my first letter. It seems to be something along the lines you are stating. I found some of my combos worked - refrencing the exact same maintenance form. So, I made every single property and method the same between one that did work and one that didn't - no effect. Then I changed the Read Only flag on my default data session ( your solution ). Some then worked but others still failed. I then examined the tables directly. Sure enough the data was there just not in the index and I noticed by cycling the Set deleted On/Off I could get the data to appear and disappear. Finaly I packed and or reindexed the files and the data appeared fine. I can't do this in the production environment as it requires exclusive access to the file which by definition I don't have.
>A this point I am again stuck.
>
>Thanks
>Terry

I read through the thread and looked at what the others have said. First, your primary keys should be absolutely unique. That is the definition of a primary key. There is no reason the user needs to see it. If you need information on creating them, look at my web site.

Second, by filtering your index tags you are basically turning off Rushmore. Again, there is an article on my web site.

Third, you should use the FLUSH command after every update (add, edit, and delete). It will make sure that the buffers on the workstation are flushed to the server.

Finally, make sure that ALL write buffering on the server's hard drives is turned off.