Allin Cottrell a écrit :

On Wed, 9 May 2007, a b wrote:

  

 Will it be possible to add, ... the number of observations actually 
in use ?
    

 Could you do the same for the <summary> command ? 
    

That has its own problems, in that the summary statistics may be 
based on different numbers of observations for ddifferent 
variables, depending on the pattern of missing observations.

That's why I think, it will be useful to know how many observations are really used for each variable. That' the reason too, some time ago I suggested 2 ways for computing the matrix correlation based on commun or pairwise observations. RATS does it, Eviews either! Why not Gretl ??? :-[
Anyway, these are stuffs you may think about in your spareeeeeeeeeeeeeest time ! :-)

A bunch of thanks,
Artur