Statalist The Stata Listserver


[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

st: RE: capture problem


From   "Nick Cox" <n.j.cox@durham.ac.uk>
To   <statalist@hsphsun2.harvard.edu>
Subject   st: RE: capture problem
Date   Thu, 7 Sep 2006 21:34:01 +0100

There is some confusion of terminology here. -arima- is 
a command which is defined by a program. From your posting 
it is being used inside a .do file (which is not a program 
in the strict Stata sense, although informally some people 
refer to writing .do files as (part of) Stata programming). 

However, I think that when you say "program" you are referring 
to the .do file you are running. So, "stop the program" is ambiguous, 
depending on whether strict Stata terminology or some 
other terminology is being used.  

Anyway, to reprise the point of -capture-: 

-capture- always sits on the outside of what it is capturing. 

It never climbs inside the command to do something 
different from what otherwise would have happened. 

If you say 

capture <command> <details> 

then 

<command> <details> 

is going to go ahead just as it would have done if you 
have never issued -capture-. The difference is what 
happens to the output. 

So, if you want to tune what -arima- does, you will have to 
tune -arima-'s <details>. 

And an error is still an error, regardless of where
the associated output goes. 

Positively, what you want may be -do, nostop-. See
also

SJ-6-2  pr0023  . . . . . . . . . . . . . . . . . .  Stata tip 32: Do not stop
        . . . . . . . . . . . . . . . . . . . . . . . . . . . .  S. P. Jenkins
        Q2/06   SJ 6(2):281                                      (no commands)
        tip for using the do command's nostop option

Nick 
n.j.cox@durham.ac.uk 

Richard Boylan
 
> In a  long .do file, I have the command:
> 
> capture arima...
> 
> However, the program stop with the message:
> 
> flat log likelihood encountered, cannot find uphill direction
> 
> I thought that this was the point of the capture command; i.e., to
> insure that the program does not stop if there is an error mesage.
> 
> Is there any way around this? I.e., to insure that the program
> continues even if in part of the program it encounters an arima
> estimation that won't converge?

*
*   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/



© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index