Notes in-line
--
Regards
Jonathan Lewis
http://www.jlcomp.demon.co.uk
The Co-operative Oracle Users' FAQ
http://www.jlcomp.demon.co.uk/faq/ind_faq.html
April 2004 Iceland http://www.index.is/oracleday.php
June 2004 UK - Optimising Oracle Seminar
"Anurag Varma" <avdbi_at_hotmail.com> wrote in message
news:VRVac.1893$KK3.1333_at_nwrddc02.gnilink.net...
>
> "VC" <boston103_at_hotmail.com> wrote in message
news:eRTac.157658$po.943948_at_attbi_s52...
>
> But here is one example which might be fixable by histogram.
Even if this one is 'fixable', you can always build a
better (worse ?) example. A histogram is limited to
a precision of 254 buckets, which means that a spike
in the data that is as large as 0.4% of the data can
be missed completely (in fact with a little luck, a spike
as large as 0.8% could be missed - it is possible to
nearly fill two buckets without appearing on two
bucket boundaries).
>
> SQL> select c2, count(*) from t1 group by c2 having count(*) > 1;
>
> C2 COUNT(*)
> ----------------------------------------- ----------
> A 5
> B 95
>
> SQL> select c2, count(*) from t2 group by c2 having count(*) > 1;
>
> C2 COUNT(*)
> ----------------------------------------- ----------
> X 99
>
> SQL> select count(*) from t1;
>
> COUNT(*)
> ----------
> 14099
>
> SQL> select count(*) from t2;
>
> COUNT(*)
> ----------
> 14099
>
Received on Thu Apr 01 2004 - 08:49:58 CST
