>>a) Even if that table has 0 records in it now you can calculate "how many records will fit in X bytes". Roughly calculation is:
>>

>>lnTargetBytes = 2*1024*1024*1024
>>? int((m.lntargetbytes-HEADER())/RECSIZE())

>>
>>If table has no memo/general/BLOB type fields thjis would give you your limits.
>
>That's great, thanks.
>
>>b) You can keep an eye on reccount() then or directly on file size itself.
>>
>>PS: Such a case sounds to be a great candidate for SQL server (or some other backend).
>
>It's a standalone app and I favour the lightweight DBFs for stuff like that so I don't want to have to install local instances of SQL Server.
>
>Kev

Yes I can understand the ease working with DBFs and 'standalone' should be simplier:) However don't forget they always tend to grow up quickly:)
Even if you start with DBFs design it in a way that switching to a backend would be simple.
Cetin