I'm using sampling to try and cut down in the time the following runstats command takes. However, it takes just about as much time with sampling compared to when I don't use sampling. Any ideas why?

DB21085I Instance "db2inst1" uses "64" bits and DB2 code release "SQL08022"
with level identifier "03030106".
Informational tokens are "DB2 v8.1.0.89", "special_15295", "U800790_15295", and
FixPak "9".
Product is installed at "/opt/IBM/db2/V8.1".

Solaris 8


runstats on table LS2USER.TLS_GL_ENTRY with distribution and sampled detailed indexes all shrlevel reference tablesample bernoulli(10) ;

How big is the table? How volatile is it? When was the last time a reorg was performed on the table?

Andy

I would only collect distribution on key columns (those that are indexed).