Hi
I have experinced a problem similar to Fred
running a long do file with several graph command (and lot of other good
stata stuf :-)
Try, while such a do-file is running, to click the read stop buttom, and
you will have to click several times in order to make stata completly stop
NOW.
Then typical you have stoped stata in the mittel of some graph call
(long interval sampling ;-)
then run the same (completly valid ) do file again
and you will get strange errors like Fred did.
a resistend attitude (=runing the do file a third time)
usual get rid of the no-real errors and produce the things you want.
I see that -discard- is a more professional appraoch ;-)
Best regard
Lars
On Fri, 3 Oct 2003, Nick Cox wrote:
> Fred Wolfe
> >
> > I was trying to make an -fpfit- graph, but had a typo
> > (fpit). Stata then gave me an error. When I corrected the
> > typo, the next (correct) command failed (see below). I then
> > added various schemes to the command and fpfit ran without
> > trouble. It would seem that Stata is leaving something
> > behind after the first failure that induces the second.
> >
> > The error messages are clearly inscrutable ...
> >
> > . twoway fpit pure plaggi_scale if luse, estcmd(logistic)
> > Super invalid member variable identifier
> > r(198);
> >
> > . twoway fpfit pure plaggi_scale if luse, estcmd(logistic)
> > invalid line in style file graphstyle/default:
> > set declared_xsize scheme x
> > invalid line in style file graphstyle/default:
> > set declared_ysize scheme y
> > invalid syntax
> > r(198);
> >
> > . twoway fpfit pure plaggi_scale if luse, estcmd(logistic)
> > scheme(s2mono)
> >
> > . twoway fpfit pure plaggi_scale if luse, estcmd(logistic)
> > scheme(s1mono)
> >
> > . twoway fpfit pure plaggi_scale if luse, estcmd(logistic)
> > scheme(s2color)
> >
> > . twoway fpfit pure plaggi_scale if luse, estcmd(logistic)
> > scheme(s1color)
> >
> > . twoway fpfit pure plaggi_scale if luse, estcmd(logistic)
> > scheme(sj)
>
> I can't reproduce this with -twoway fpit <stuff>-. That
> for me produces a very reasonable message.
>
> Bug or not, Stata can't read your mind here. Next time, type
>
> . discard
>
> to flush out all the bits and pieces loaded and
> allow it a clean start on your next command.
>
> As often mentioned, the graphics programs are
> class-based. Mostly that's a real feature. It also
> means that on occasion there is a string
> of messages several deep as you get booted out
> of a series of nested programs. Admittedly,
> perhaps three people at Stata Corp understand
> most or all such messages, but usually
> you can do your own debugging by identifying
> the typo responsible.
>
> Nick
> [email protected]
>
............................................
: Lars Korsholm Ph.D., :
: Department of Statistics :
: University of Southern Denmark :
: Sdr. Boulevard 23A, 2. :
: DK-5000 Odense C :
: Phone +45 6550 3608 :
: Mobil +45 3033 3617 :
: Fax +45 6595 7766 :
: Mail [email protected] :
: Web http://www.isd.sdu.dk/~korsholm :
:..........................................:
*
* 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/
