-------------------
> Nick,
>
> I was sloppy in the e-mail, not in the original code. Also,
initially I used
>
>
> forvalues i=1/`x'
>
> to no avail. So I switched to the programming manual's less elegant
>
> forvalues i=1(1)`x'
>
> out of desperation.
>
> Anyway, here's a strange ambiguity:
>
> Suppose mytempfile has 17 observations. The code
>
> use "`mytempfile'"
> display _N
> local x _N
> display `x'
>
> will have the same effect as
>
> use "`mytempfile'"
> display _N
> local x=_N
> display `x'
>
> The effect is that the display command will twice show the right
number, 17,
> in both cases.
>
> The difference between two slabs of code is that
>
> forvalues i=1/`x'
>
> does not work after the first, but works fine after the second.
>
> In other words, the equal sign makes no difference to the value of
the local
> variable x, yet it does matter to forvalues.
>
_N is strange. It is not a variable, or a macro. There are, however,
contexts in which it is evaluated automatically, and in effect it
behaves in those contexts as it were a macro -- or even a variable.
That is, Stata sees _N and replaces it by the appropriate (or
numbers!).
-display- is one example. -display- is prepared to do some
work for you, and arguments to -display- which are expressions are
evaluated, Indeed -display- has a major role as a calculator.
-forvalues- is at the other extreme. _N is just not
accepted as an argument, as you found.
Yet another way to do it in Stata 8 is
forval i = 1 / `c(N)' {
....
}
Nick
[email protected]
*
* 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/
