Re: Whatcha' wanta have?????
In a feature-wish thread, Madison Pruet said a logged or unlogged TRUNCATE-style feature also exists in DB2. Neil Truby asked that it not become an us-versus-them argument. Madison said he was only pointing out to Jonathan Leffler that the functionality might be needed. Andrew Hamm joked that Jonathan should sign off if he got the logged TRUNCATE. No decision.
Auto-generated by Claude from the posts below — may be imperfect; read the full thread.
"Jonathan Leffler" <jleffler@earthlink.net> wrote in message news:402EEA68.9010107@earthlink.net... > Serge Rielau wrote: > > Neil Truby wrote: > >> "Serge Rielau" <srielau@ca.eye-be-em.com> wrote: > >>> Andrew Hamm wrote: > >>>> Mark Denham wrote: > >>>>> Some indexing additions: > >>>>> > >>>>> 1) Inclusion of columns in the index that are not part of the index > >>>>> key but help speed processing. This feature is available in a number > >>>>> of other rdbms's. > >>>> > >>>> Waaaaaa? Please explain? > >>> > >>> This feature applies to unique indexes. > >>> Let's presume you have an employee table. The PK is empno. > >>> You want to optimize lookups of employee names by empno. > >>> Without this feature the fastest to do this (assuming a non trivial > >>> table size) is to fetch the rowid from the index and then do an fetch by > >>> rowid from the table to get the name. > >>> If you INCUDE empname in the unique index then you save the the table > >>> access. [...strike incorrect sentence...] > > I don't see that it has to apply solely to unique indexes. It > basically allows a key-only scan to also pick up the other critical > data without having to fetch the data page. So, if you've got a table > of zip-codes and state codes, you index on the zip-code - which > probably needs to be unique - and include the state code too, and you > never have to go to the table which contains demographic data stored > in blobs for the state code. > > The Orrible database has a feature along these lines, I believe. This is also a feature of DB2..
"Madison Pruet" <mpruet@comcast.net> wrote in message news:rmOXb.324282$na.477931@attbi_s04... > > "Jonathan Leffler" <jleffler@earthlink.net> wrote in message > news:402EEA68.9010107@earthlink.net... > > Serge Rielau wrote: > > > Neil Truby wrote: > > >> "Serge Rielau" <srielau@ca.eye-be-em.com> wrote: > > >>> Andrew Hamm wrote: > > >>>> Mark Denham wrote: > > This is also a feature of DB2.. Madison, could we try to avoid turning this thread into an "us-vrs-them" thread? :-)
Ouch! ;-) Actually I was pointing out to Jonathan that we might need this functionality. ;-) M.P. "Neil Truby" <neil.truby@ardenta.com> wrote in message news:c0ohod$18dn12$1@ID-162943.news.uni-berlin.de... > "Madison Pruet" <mpruet@comcast.net> wrote in message > news:rmOXb.324282$na.477931@attbi_s04... > > > > "Jonathan Leffler" <jleffler@earthlink.net> wrote in message > > news:402EEA68.9010107@earthlink.net... > > > Serge Rielau wrote: > > > > Neil Truby wrote: > > > >> "Serge Rielau" <srielau@ca.eye-be-em.com> wrote: > > > >>> Andrew Hamm wrote: > > > >>>> Mark Denham wrote: > > > > This is also a feature of DB2.. > > Madison, could we try to avoid turning this thread into an "us-vrs-them" > thread? :-) > >
Madison Pruet wrote: > Ouch! ;-) > > Actually I was pointing out to Jonathan that we might need this > functionality. ;-) How's about he signs off on that if you can make the other members of the planning committee give him the logged TRUNCATE?