seq scan counts

I have been running the following: select first 20 dbsname, tabname, seqscans from sysptprof where seqscans > 0 order by 3 desc; AIX 4.2.1 IDS 7.30.UC7-1 F50 machine One of my larger tables is showing a large number of sequential scans since we upgraded (from 7.24.UC1). I have been tracking user sql's and jobs (ps). It appears to me that when the table is joined to other tables, that if any table in the query is sequentially scanned, that it is also incrementing the seqscans counter for the large table too. I have isolated some of the code and sql and run with set explain on, and it does not seq scan the large table, just the smaller join tables. The onstat -p shows 680k seqscans, while the select sum(seqscans) shows only 580k seqscans. I do not remember this happening in the old version. Can anyone enlighten me? Chris

Chris Burton wrote: > > I have been running the following: > > select first 20 dbsname, tabname, seqscans > from sysptprof > where seqscans > 0 > order by 3 desc;> > AIX 4.2.1 > IDS 7.30.UC7-1 > F50 machine > > One of my larger tables is showing a large number of sequential scans > since we upgraded (from 7.24.UC1). I have been tracking user sql's > and jobs (ps). It appears to me that when the table is joined to > other tables, that if any table in the query is sequentially scanned, > that it is also incrementing the seqscans counter for the large table > too. I have isolated some of the code and sql and run with set > explain on, and it does not seq scan the large table, just the smaller > join tables. > > The onstat -p shows 680k seqscans, while the select sum(seqscans) > shows only 580k seqscans. > > I do not remember this happening in the old version. > There was a thread on this in early March. V7.24 reporting bug. Check the IIUG archive of the group for details. Art S. Kagel