st: Re: triangular matrices

[Kit Baum <baum@bc.edu>]

Elementary, my dear Brendan. No need to do anything in the Stata  
matrix language in order to get the Stata variable into Mata. If you  
want to do anything with matrices, you should pick up some  
familiarity with Mata; it makes life a lot simpler.

clear
input v1 v2 v3
1 1 0
1 2 12
1 3 13
1 4 14
2 2 0
2 3 23
2 4 24
3 3 0
3 4 34
4 4 0
end
mata:
st_view(a=., ., "v3")
st_matrix("b",invvech(a))
end
mat list b



Kit Baum, Boston College Economics
http://ideas.repec.org/e/pba1.html
An Introduction to Modern Econometrics Using Stata:
http://www.stata-press.com/books/imeus.html


On Nov 18, 2006, at 2:33 AM, Brendan wrote:


> Thanks for that solution -- it points me towards Mata which I had
> been studiously ignoring.
>
> Unfortunately, if I understand things correctly, this introduces a
> size limitation: for an m*m dissimilarity matrix, the "mkmat
> v3,mat(a)" statement will create an m*(m-1)/2 vector. If I want m
> to run up close to the "set matsize" limit, this is not possible.
> (My understanding might be wrong: if I say "set matsize 100" I can
> create 100*100 matrices but not 101*1 vectors?).
*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/