

Archives of the TeradataForum
Message Posted: Wed, 07 Jan 2004 @ 17:13:22 GMT
Subj:   Re: Intervals for Statistics 

From:   Victor Sokovin 
Correction of the typo:
On Wed, 7 Jan 2004 12:01:36 0500, Victor Sokovin wrote:
 First the frequency distribution is calculated. Based on it TD knows the number of loners = values whose frequency is greater than or equal
to 1/200 x (base table cardinality). The allocation of intervals starts with the loners and highbiased intervals. By design there can be at most
two values in a highbiased interval (1 or 2 loners), and the total number of highbiased intervals is 99.  
should read
First the frequency distribution is calculated. Based on it TD knows the number of loners = values whose frequency is greater than or
equal to 1/200 x (base table cardinality). The allocation of intervals starts with the loners and highbiased intervals. By design there can be at
most two values in a highbiased interval (1 or 2 loners), and the *maximum* number of highbiased intervals is 99.
Victor
 