Re: st: Speed with large panel datasets

[Gordon Hughes <G.A.Hughes@ed.ac.uk>]

This throws up something that I find even odder.  I investigated 
Maarten's suggestion of using -statsby-.  The Stata help information 
even has an example of -statsby- combined with -arima- as in:

webuse grunfeld, clear
tsset company year
statsby _b _se, basepop(company==1) by(company):  arima invest mvalue 
kstock, ar(1)

So, my test was:

use xxx.dta, clear
tsset panel time, monthly
<various Stata commands to create variables>
save temp.dta, replace
statsby _b _se, basepop(panel==1) by panel: arima ....

Stata's response

no; data in memory would be lost
r(4);

I can get round this by adding

use temp.dta, clear
tsset panel time, monthly

after the -save- and before the -statsby- but what is the logic, 
given that I have already used -save-?

Incidentally, -statsby- is nothing like as fast as the -reshape- approach.

Gordon Hughes
g.a.hughes@ed.ac.uk

*
*   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/