I have a file that uses a symbolic field as a Btree indexed field and the field only contains two values. The file has approximately 300,000 records in it, does anyone know of any limitations or problems with using this field as an index?

With only two values, you don't gain all that much of the power of indexing (but I don't think anything breaks.) Sometimes a little bit of redesign can gain an enormous amount of power. As a small example: instead of an boolean paid or unpaid for receivable invoices, index the date paid field. That way you don't have 150,000 records in each of the two values, and the nulls (the ones of interest) are faster and easier to find.

The only thing I know about indexing on a symbolic is that the indexes don't automatically get updated when the values in columns they point to get changed. You have to rebuild your indexes on that symbolic to make sure they are up to date. At least that is what we have to do on the one symbolic that we have indexed in one of our applications. Out of the four applications, we only have the one symbolic indexed.

I hope this helps.

Michael Slack

View this thread on the forum...