Hi Bruce,

I'm thinking this was your 'theory of lost data' or something like that that you are referring to. Is that correct?

If so, I don't remember how you solved it.

But in the context of this problem how would you rate this idea. . . SET REFRESH???

As usual the Help notes that REFRESH applies only to BROWSE. But we know differently.
Now the Help also says that 0 is the default, meaning that no records will ever be checked for updates by something else.
Assuming that both the client app here and the server app here both have REFRESH set to default, I *could* see something like this (and yours) happening **IF** REFRESH applies to the sequence in question.
I guess that there is no way to confirm this (at least from docs) and the condition sounds hard to duplicate for sure.

I'm just thinking that setting REFRESH would be at least as useful as doing a FLUSH after each TableUpdate.

Any thoughts

Jim N

>>Hi Tim,
>>
>>Thanks for the added detail. Not sure I'll be much help, but hopefully others can be.
>>
>>I've not had the pleasure of a Win95 client but recollect mention of some problems there, more to do with locking I think (than "lost" records).
>>
>>I have heard of strange things happening on NT Server when heavy printing is going on. Things like random 'not finding/opening of index file' yet no error regarding (problem found later when index use attempted). There were other anomalies too. It was Barbara Paltiel (often on UT) who experienced these oddities and I recollect that she got around most by getting agreement to keep printing low while app ran.
>>
>>It sure wouldn't hurt to code the alias in your TableUpdate command, just to be certain of things. I also assume that you are checking the return from TableUpdate to confirm success.
>
>Just looking over this thread...this reminds me a little of the recent problem I had where sometimes an entire server table was pulled onto the local machine, and all the Tableupdates were done locally. Then there was a network problem, and the server table never got updated with anything...
>
>Does this bear any resemblence?