The situations are quite distinct.
1. There are no observations [in any sense] for which -rep78 == 0-, so
the core code of -summarize- never gets called up.
2. When -summarize- reports "Obs 0" for -rep78 == .- this is to be
understood as a contraction of "Observations for which there are
non-missing values 0". There is hardly enough space in the Results
window even to abbreviate that.
The point is explicit in the manual entry, but the definition of r(N) in
the online help could be tweaked to make it more obvious still.
Nick
[email protected]
Martin Weiss
" Interestingly,
. su rep78 if rep78==0, detail"
But that is not the surprising part, since there are no "0" values for
"rep78". The weird thing is that
*************
su rep78 if rep78==.,d
*************
returns "no observations" whereas -su rep78 if rep78==.- reports 0
observations.
Kit Baum
On Feb 11, 2010, at 2:33 AM, N.J. Cox wrote:
> - -summarize-
> ===========
>
> . sysuse auto
> (1978 Automobile Data)
>
> . su rep78 if rep78 == .
>
> Variable | Obs Mean Std. Dev. Min Max
> -
-------------+--------------------------------------------------------
> rep78 | 0
>
> . ret li
>
> scalars:
> r(N) = 0
> r(sum_w) = 0
> r(sum) = 0
Interestingly,
. su rep78 if rep78==0, detail
Repair Record 1978
-------------------------------------------------------------
no observations
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
