st: Re: invalid file specification

[Kit Baum <baum@bc.edu>]

In this case a useful technique (which is easily disabled when you  
get tired of seeing it) is just to put a line

display "Now reading `bestand'"

just before the insheet, which would make it obvious which file it is  
processing (if any error conditions occur within that file, for  
instance) but would also show you that in fact the local had not been  
given a value.

It is true that Stata does not make it easy to determine which line  
is giving you grief in a loop context, but as Austin suggests good  
old-fashioned debugging techniques work well. Putting output lines  
into the do-file and just commenting them out when you no longer need  
them is one of those techniques.


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


On Jul 8, 2007, at 2:33 AM, statalist-digest wrote:


> Hi Austin,
>
> > In these situations, if you -set trace on- you can usually see the
> > source of the problem immediately.  In this case, I suspect you have
> > forgotten that -if exp {- must appear on a line by itself (see -help
> > ifcmd-), i.e. you must have a delimiter between "{" and "local"  
> > below.
> Ok, thank you. I couldn't think that that was the problem from an  
> "invalid
> file specification" error, so thanks for the help.
>
>
> > Your postscript assumes either Stata (or you) would want to -set tr
> > on- all the time, I think.  Or somehow magically reach into the guts
> > of programs and tell you where you forgot a keystroke.  Neither is
> > reasonable IMHO.
> If stata can detect a syntax error, it does know what was not along  
> the
> rules of the syntax and where it happened. It can tell me that  
> along with
> the error message that it produces now. "Syntax error"  is not at all
> informative. Of course I made an error, otherwise the code would run.
>
> - -set trace on- does work, except that it produces a lot of  
> output. (Yes, one
> can set depth.) But it does include the syntax error that actually  
> made my
> code fail. If only it would output that error and the code that  
> triggered
> it, nothing more, certainly not error message that only confuse.
>
> But anyway, thanks for your help. If my code fails again next time,  
> I will
> - -set trace on- and hope that I can spot the true error message.
*
*   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/