Re: Large logicbase - how to deal with it?

how does one "swap in" and "swap out" parts of the
logicbase to deal with a query that might be out on disk?

Amend that to read "to deal with a query which needs something which
might be out on disk?" Say you want an ancestor or descendant of
someone, and either the "someone" is out on disk, or an ancestor or
descendant is out on disk. This is assuming, of course, that one
could break up a logicbase in such a manner in the first place.

How does one manage a large logicbase of facts?

.

Relevant Pages

  • Re: Large logicbase - how to deal with it?
    ... logicbase to deal with a query that might be out on disk? ... could break up a logicbase in such a manner in the first place. ... How does one manage a large logicbase of facts? ...
    (comp.lang.prolog)
  • RE: Has any Informix DBA had to do the following?
    ... I've never had that problem myself using IDS, granted then I was using c and esql/c for the connections so I knew every piece of code and had no thread issues. ... Disk access - busy disks but no predominant ... duration of the query. ... Has any Informix DBA had to do the following? ...
    (comp.databases.informix)
  • Re: Single-Threading / Performance issues
    ... Apparently not since at least 3 physical disks are required for RAID-5. ... This is normal but the percent disk utilization metric isn't much use ... might want to set SQL Server 'max server memory' if you have other apps ... > From Query Analyser, I run my import proc - which first copies data ...
    (microsoft.public.sqlserver.server)
  • Re: Whats the deal with PAGEIOLATCH_SH?
    ... In addition to what Greg suggests are you monitoring the Checkpoint Pages ... Checkpoints can really hose a disk with low ... About 20 minutes into the query everything goes bad. ... > I'll work on getting the real execution plan, ...
    (microsoft.public.sqlserver.programming)
  • Re: Whats the deal with PAGEIOLATCH_SH?
    ... About 20 minutes into the query everything goes bad. ... > That sounds like a disk I/O problem to me, ... > I'll work on getting the real execution plan, ... >>scans or inefficient joins such as hash joins. ...
    (microsoft.public.sqlserver.programming)