Re: Real world issue:- OT recreational interval
Date: 19 Sep 2006 00:17:16 -0700
Message-ID: <1158650236.533726.39590_at_m73g2000cwd.googlegroups.com>
Bob Badour ha scritto:
Count is not idempotent, but in this case it does not operate
> on any duplication. (Replication means something different in the
> database field.)
>
I also agree (can't believe I agree with Bob :) that for terminology
it's better to avoid
"replication" word in favor of "duplication" (if you see a better
term please **notice**)
If idempotent is not the term. What is a good term?
quoted from:
http://groups.google.it/group/sci.math/browse_frm/thread/17781c0124345f6e?scoring=d&hl=it
William Hughes
>More important perhaps, CountDistinct(2,2) = 1
>It is clear that "replication insensitive" cannot be the same as
>idempotent.
>For one thing, any constant function that can be replication
>insensitive
>(the definition given is not complete) is. No idempotent function can
>be constant.
http://groups.google.it/group/comp.databases.theory/browse_frm/thread/538735631f141027?hl=it
Marshall
>cd(cd(5,6),2)
...
>Consider how you would use + to make sum:
>(5+6)+2
>or binary min to make aggregate min:
>min(min(5,6),2)
>Extra credit:
>avg() is one aggregate that doesn't seem to fit with
>the rest. Why is that?
3 questions:
Now let's proceed further
when you say:
1
let me understand, are you implying that you could
compute the aggregate function recursively ?
2
Would you please explain how do you compute the
following "aggregate" functions for n= 3:
- range of values (difference between max and min)
- median
- standard deviation
3.
would you please provide the "binary" (n=2)
version of the above listed functions
-P Received on Tue Sep 19 2006 - 09:17:16 CEST