Statalist


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

st: AW: Tracing in Stata


From   "Martin Weiss" <[email protected]>
To   <[email protected]>
Subject   st: AW: Tracing in Stata
Date   Tue, 5 May 2009 18:49:31 +0200

<> 

In such cases, I either look at the error very closely and try to figure out
which command could possibly have caused it - most of the time that narrows
it down quite substantially. Or I let my template below record the entire
-trace- with -tracedepth- at infinity (grab a cup of tea in the meantime),
have it turned into a pdf and search it in the pdf reader for the offending
command...


*************
//to log in an additional log and have it transferred to pdf

//can set -linesize-
set linesize 120

//close if open
capt log close newlog
di in red _rc

log using mylog.smcl, name(newlog) replace
//do whatever you want in the log

//trace the thing with Ben Jann`s tool
//tr: 

//or use conventional set trace on/off
//set trace on

//set trace off

log close newlog

//translate
translate mylog.smcl mylog.ps, replace
//need Miktex utilities running on computer for this step
!ps2pdf mylog.ps
capt erase mylog.ps
!start mylog.pdf
*************



HTH
Martin


-----Ursprüngliche Nachricht-----
Von: [email protected]
[mailto:[email protected]] Im Auftrag von John Bates
Gesendet: Dienstag, 5. Mai 2009 18:41
An: Stata
Betreff: st: Tracing in Stata

<>

Hi,

Is there any way to figure what line in a for loop Stata is erroring on?
The best way I know of is to do -set tracedepth 1- and -set trace on-.  This
works, BUT it's annoying because Stata traces down into function calls that
I don't care about.  For example, assume there is an error somewhere in the
following (potentially very long) for loop:

forval x = 1/20 {
          regress...
          probit....
       ...
}

I just want to know which line in the for loop the error is without having
trace go down into the regress ADO code etc.  This is not a problem outside
of the for loop because Stata goes line-by-line down do files by default.
All in all, a very minor issue, but one that frustrates me.  Ideally there
would be a -set tracedepth 0- option which would address this.

Lirac



      

*
*   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/


*
*   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/



© Copyright 1996–2024 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index