Hi Rick,
Check out the following:

SYS(2021)
CANDIDATE()

As for unique, I wouldn't recommend using it for indexes. Unique simply filters out duplicate entries in the index. It does not enforce uniqueness in the table itself. Except for counting, I can't think of a reason to use it.

FYI, the index bloat issue was discussed last week (or two weeks ago) and it was deemed to be a non-issue in the more recent versions of VFP. Jim Nelson started the thread if you want to search for it.

HTH.

>Hello world,
>
>I would like to get all the information on the indexes in a database so I can recreate the indexes from scratch.
>
>I have two purposes, 1) thoroughly "reindex" the tables, and 2) back up the tables without the indexes, and then rebuild the indexes in the event of a restore. Concerning 1), Menachem Bazian has written that CDX files can bloat, and this can only be remedied with DELETE TAG ALL, after which one must reconstruct the persistent relations in the database. Concerning 2), I have a back up bottleneck, and it would be convenient to eliminate the CDX files from the back up set.
>
>I can get the tag names and kex expressions from the TAG() and KEY() functions. I can get the primary key tag with dbGetProp(). How do I get the order (ascending or descending), the index type (candidate, unique, regular) for non-primary tags, and whether there is a filter?
>
>Any general comments, suggestions or caveats conerning this approach?
>
>Thanks,
>
>Rick Graves (displaced American currently residing in Hong Kong)