Patrick,

Here is some information that I put together from the ALLBASE/SQL DBA guide.

A clustering index attempts to place rows with similar key values
physically close to each other on disk. I/O overhead is therefore reduced
and performance improves especially for retrievals that fetch rows in
sequential order based on the key value.

The cluster count of an index indicates how many times ALLBASE/SQL has to
access a different data page to retrieve the next row during an index scan
or how well the data of an index are sorted. The efficiency of clustering
is best if CCOUNT (cluster count) is equal to the NPAGES (number of index
pages) and worst if CCOUNT is equal to NROWS (number of rows).

Dropping and recreating the indexes would definitely help but I am not
sure of the percentage of gain. For optimal efficiency on tables using the
clustering index it is therefore recommended to unload and reload the
table. The specific steps to accomplish the same are found on page 7-17
of the Allbase/SQL DBA manual.

Hope this helps.

Regards,
ashish
CSY-DB

____________________________
email: [log in to unmask]
voice: TN or (408) 447-7820
____________________________