The following is a school example (awful english?) for using histograms.
We have a table with 200.000 rows. The table is very often indexed.
Let's have a look of two indexes. The first index have 4 distinct values
for 200.000 rows. But there's one value that is only about 100 times
in the whole table. Ok, we make the histogram and the CBO uses the
index. Easygoing, we will wait for the future.
Unfortunately, there is a second question which use the second index too.
This index with no histogram but better overall distribution now
gets the CBO's attention. But this calculation seems to only base on the
overall distribution. In the special case the value for the first index
is much more distinctive than for th second index.
Has the CBO forget the information out of the
histogram?
-----== Posted via Deja News, The Leader in Internet Discussion ==-----
http://www.dejanews.com/ Now offering spam-free web-based newsreading
Received on Sat May 23 1998 - 01:35:07 CDT
