I think you were right first time, i.e. the details of what is inside
the macro do not seem important in explaining your problem.
Nick
[email protected]
Gabi Huiber
It's just a set of variable names, like this:
global dates "last_ar_payment next_bill curr_rate_exp strt stop"
The offending loop then does this, in the same do-file:
******* ******* put dates in Stata format
foreach k in ${dates} {
gen `k'=date(substr(`k'_date,1,10),"mdy")
drop `k'_date
rename `k' `k'_date
format `k'_date %d
}
Then in the second do-file I am trying to do this:
foreach k in $dates {
gen `k'_sunday=`k'_date-dow(`k'_date)
drop `k'_date
format `k'_sunday %d
}
But the curly bracket syntax doesn't work anymore. I need to use
$dates like above.
On 6/6/08, Nick Cox <[email protected]> wrote:
> You are omitting some details that might be the key to the question.
> What is in this problematic global?
>
Gabi Huiber
>
> I got into the habit that I use the ${globalmacro} syntax instead of
> the plain $globalmacro, on the chance that I sometimes have to
combine
> such macro names within other macro or variable names.
>
> But this syntax sometimes elicits an error message from Stata, like
so:
>
> foreach k in ${dates} {
> program error: code follows on the same line as open brace
> r(198);
>
> At other times, Stata doesn't mind. In this particular example I
> declared the ${dates} macro in one do-file, and used it with foreach
> as shown above within the same do-file without a problem.
>
> Then another do-file called and executed the first do-file, and later
> attempted to make use of the same ${dates} macro in another foreach
> construct. It was this second time that I got the error response
> above.
>
> The syntax $dates in this second foreach loop is accepted and all is
> well. But I can't find any reason for this inconsistency. Both
> syntaxes should always work.
*
* 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/
