I need to take already summarized data in the logs, aggregate it from a large group of servers, and build an si-type index. Looking at si-generated data from sistas fields, I have deduced the following meanings but need further clarification
psrsvd_ct_FIELDNAME = count
psrsvd_nc_FIELDNAME = Also Count?
psrsvd_sm_FIELDNAME = sum
psrsvd_ss_FIELDNAME = sum of squars
psrsvd_vt_cnt = ?? some kind of variance ??
So is ct = count, what is nc really for, what formula do you use for SS (does it include std-dev or is it a simple sum of squares), and what is vt?
Thanks,
Blaine
nc = numerical count (number of values for this field that can be interpreted as numbers), e.g. computing the average would be sum/nc
ss is just simple sum of squares, i.e. X1*X1 + X2*X2 + X3*X3
vt is actually originally stood for "valuetype", but what we store in it is actually just the maximum precision of the numerical values of the field. This is so that if you average together a bunch of values like 9.5,10.5,11.5, you get 10.5 and not 10.500000 or 11
Hi, I am just catching up on you your answers followed by comments...
What does "psrsvd" as in psrsvd_* stand for here ?
nc = numerical count (number of values for this field that can be interpreted as numbers), e.g. computing the average would be sum/nc
ss is just simple sum of squares, i.e. X1*X1 + X2*X2 + X3*X3
vt is actually originally stood for "valuetype", but what we store in it is actually just the maximum precision of the numerical values of the field. This is so that if you average together a bunch of values like 9.5,10.5,11.5, you get 10.5 and not 10.500000 or 11
nc is numerical count, i.e. the number of values that can be entirely as numbers, where as ct is the total count.
I.e. if your field had the following values:
0,foo,1,bar
ct=4 and nc=2
is there a difference between ct and nc