Thanks Jim! The only additional comment I would make here is that the information on the index is pulled from OUR tables HERE and not at the client's site. That information is then put into a procedure file that we can run at the clients site to recreate indices when necessary. In everything below I agree 100% and would prefer a metadata approach, but once again, no matter what I recommend or how I substantiate any recommendation I make, I am always forced to do as the 'powers that be' wish. If I was working alone however.... I could do what I prefer. :0)
Tracy

>Tracy,
>
>My comment is still pertinent. If the index file's header record is corrupt your code will not be able to determine the information about the index. You would be bettere served to create a dbf that stores the table adn index information (like SDT does and use that to recreate everything). Also with that approach you would only need to update your data dictionary and run a database update routine to add or remove fields from tables and to update index files.
>
>Using the functions to read the information from an existing index can be frustrated by two things, a corrupt index header or a missing index file. Using a meta data approach is not affected by either of these.