Re: The TransRelational Model: Performance Concerns
Date: Fri, 12 Nov 2004 00:46:10 GMT
Message-ID: <mJTkd.22827$V41.7166_at_attbi_s52>
"Mikito Harakiri" <mikharakiri_at_iahu.com> wrote in message news:IBTkd.14$vi7.133_at_news.oracle.com...
> "Marshall Spight" <mspight_at_dnai.com> wrote in message
> news:K7Tkd.396968$D%.165120_at_attbi_s51...
> > "Josh Hewitt" <lajos.nagy_at_gmail.com> wrote in message
> news:1c92edeb.0411102354.45157060_at_posting.google.com...
> > I note that if all columns are sorted, then
> > every join can be a merge join, which is darned fast. O(m+n).
>
> Merge join (and performance-wise similar HJ) is a mediocre method. It's
> better than [unindexed] Nested Loops, but all [OLTP] applications are
> usually tuned to leverage indexed Nested Loops, which is the fastest of all.
Do you have any suggestions for places to start learning about query optimization and query execution? I have gone looking but not found much of anything. But it's clearly a field that *someone* knows a lot about, just not me.
What I'd love to see is a comparison of various data structures for storing relations, including performance of various operations (insert, delete, join, etc.)
Marshall Received on Fri Nov 12 2004 - 01:46:10 CET