Statalist The Stata Listserver


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

RE: st: RE: -save- with emptyok


From   "Nick Cox" <n.j.cox@durham.ac.uk>
To   <statalist@hsphsun2.harvard.edu>
Subject   RE: st: RE: -save- with emptyok
Date   Tue, 13 Feb 2007 11:44:19 -0000

I have some sympathy with this call, echoed by Jeph Herrin, 
but not total sympathy. 

In practice the work-around is so easy that I would much 
rather the good folks at StataCorp implement many other
things first. 

Any way, what is the underlying problem? You want to
join several data files, observation-wise, and kicking
off the process is a little tricky because you have
to -use- before you can -append-? Isn't Roger Newson's
-dsconcat- one of several solutions to that? 

More crucially, this looks like a little deal from your 
point of view, but you are asking StataCorp to 
change one or more fundamental commands. 

I am not clear whether you want a change to -save- or to
-append-, but it is the same story either way. 

Changing the behaviour of a fundamental command is scary
stuff when you want is not a bug fix (which by definition
is desirable). Who knows what other commands, programs or do files might
be broken by any such change? I have no idea, but unless
you can be certain that there are no side-effects, this 
is questionable. Of course, it is StataCorp's job to 
worry about that; and be assured, they would worry about 
that. 

Even adding an extra option looks like questionable complexity. 

Stata users often forget that just adding a little feature to
official Stata can be a much bigger deal than just tweaking the code. In 
general, you need to make a hit on the help, the manuals, the 
dialogs, and the test scripts; and there are downstream implications
for maintenance, tech support and marketing. Now don't feel
sorry for StataCorp: they take your money so that they can do
this, among other things. But these changes can be a bigger deal 
than you think. 

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

Partha Deb
 
> Thanks Nick.  I implemented a solution, much like the one you 
> suggest, that 
> does not require emptyok.  But it seems to me that emptyok is 
> an "empty" 
> option if variables/observations need to be specified anyway 
> before -save , 
> emptyok- plays nicely with -append- or -merge-.  Perhaps this 
> is something for 
> the good folks at Statacorp to chew on.
> 
> cheers,
> 
> Partha
> 
> 
> Nick Cox wrote:
> > The issue here is not with -save-, but with -append-. 
> > 
> > -append- needs something more than almost nothing 
> > to join on to. I think you need to -save- your dataset
> > with variables (even though there are no observations). 
> > Fortunately, a token variable is sufficient (but evidently 
> > necessary). 
> > 
> > . clear
> > . gen completeanduttergarbage = . 
> > . save empty, emptyok 
> > . append using auto
> > . drop completeanduttergarbage 
> > 
> > Admittedly, the error message here does not indicate
> > the problem. 
> > 
> > Nick 
> > n.j.cox@durham.ac.uk 
> > 
> > Partha Deb
> >  
> >> I'm having trouble getting -save- to work with emptyok in the 
> >> way I think it 
> >> should.  Here's an example.
> >>
> >> . save temp, emptyok
> >> (note: dataset contains 0 observations)
> >> file temp.dta saved
> >>
> >> . use temp.dta
> >>
> >> . append using auto.dta
> >> no dataset in use
> >> r(3);
> >>
> >>
> >> Shouldn't I be able to append?  What am I doing wrong?  I'm 
> >> using Stata 9.2. 

*
*   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