In article <[log in to unmask]>,, #Rudi Grom (aka rudigrom)
says...
>
>As we realized in our department there is a problem with a master-dataset :
>More than 20% of all entries belong to entries where the hashing-algorithm
had
>to calculate a secondary key-entry. Due to that the performance is very bad.
>The master-dataset is 50% full. Is there a way to optimize the
dataset-filling
>? How does the hashing-algorithm work ? (I know there must be an article
>describing it )
>
 
20% secondaries is not a big problem. 40% is bad, but 20% is alright.
The IMAGE hasning algorithm cannot guarantee that there will not be
significant numbers of secondaries at any given (realistic) dataset
capacity. I guess if your dset was 1% full, you would eliminate
secondaries, but that's not realistic.
 
The hashing algorithm is described in "Primary Address Calculation" in
the TurboIMAGE/XL Database Management System. Different algorithms are
used depending on what data type (X, I, J, U, etc) is being used for
the key value. It gives general guidelines to follow to reduce the
possibility of excessive secondaries.
 
BTW, how did you know you have 20% secondaries? If you have HOWMESSY
from Robelle there is a statistic called Max Blocks which is the measure
of how far the farthest secondary is away from its primary. It directly
correlates to how many physical I/Os will be necessary to retrieve that
secondary. I once had a large master whose Max Blocks was over 9200!!!!!
Needless to say we resized immediately. A Max Blocks of 5 or less is
completely livable.
 
IN any case, you may want to explore other reasons your database or
application seems slow. 20% secondaries should not cause a huge problem,
at least in my experience.
 
 
 
>Regards Rudi
>
 
--
Mark Landin
System Manager
Superstar Satellite Entertainment
[log in to unmask]