>Perhaps you might give me guidance on this one question. It is:
>
>Specifically what, other than 1) tags, 2) keys, and 3) collation sequence, are documented factors in revealed SYS(3054) showplans' differences?
>
>We are still unsuccessful in solving this issue.
>
>1) LIST STATUS indicates that, in both instances, all of the tags and their collation sequences are identical.
>2) Both .DBCs, "development" and copy of "production" data, are on the same server.
>3) Both have close to the same number of records.
>4) The SELECT - SQL expression ran adequately for years and is unchanged.
>5) The "development" data completes in approximately 0.015 seconds
> (as did the production data until last week).
>6) The "production" data completes in approximately 23.6 seconds
>
Have you recreated the index tags from scratch to rule out corruption?
Tamar