>Some more thoughts. The reorganize option could work like this? Programmatically create a database .dbc with an obsure name in the production folder. Create all the required tables and indexes, repeating the obsure naming. Append FROM the corrupted production tables (every table, excluding the .dbc) into the obscure tables. Delete the production tables, indexes and .dbc. Rename the obscure tables to the production names (or) programmatically recreate the production .dbc and tables then append FROM the obscure tables. Upon completion delete the obscure tables?
>
>Dallas

Renaming tables that are contained in a database might be quite problematic, so I would suggest a separate folder.

The other point is, to create required tables and indices programmatically. The real challenge, I think, would be to create RI. I am not sure whether there is a way to do this programmatically yet.

Anyway, it should be easier to have a clean copy of the database, created manually. If you don't want them easily accesible, you might hide the files somewhere - in a ZIP file, or in memo fields of a table, for instance - and extract them when required.

Similar for updating data structures (new tables, new fields or changes in field lengths). You might create the ALTER TABLE commands, but a generic solution with APPEND FROM (into empty tables, with a new structure) is much easier to handle.