Danielle H. Ferry
> > I would like to run the following loop, where I have a series
> > of variables
> > produced by a regression containing an interaction of
> > state*popnLn*ttrend1.
> > "popnLn" & "ttrend1" are continuous variables, but "state" is
> > a categorical
> > variable. Since the coding for "state" has irregular spacings
> > (1, 5, 16, 20,
> > 23, etc.) the loop fails. Is there a way to make it continue
> > onto the next
> > loop if, for example, a value of i=2 is not found?
> >
> > forvalues i = 5(1)54{
> >
> > test _Ist`i'XpoXtt
> >
> > scalar f`i' = r(F)
> > scalar pr`i' = r(p)
> >
> > }
Nick Winter
> Two options. First, only loop over the values you want:
>
> foreach i in 1 5 16 20 23 {
> <<do something>>
> }
>
> Second, -capture- the attempt to test:
>
> forvalues i = 5(1)54 {
> capture test _Ist`i'XpoXtt
> if !_rc {
> scalar f`i' = r(F)
> scalar pr`i' = r(p)
> local list "`list' `i'"
> }
> }
>
> This code also leaves behind the macro list, which includes
> a list of
> the valid codes, in case you need them for subsequent processing.
Two footnotes to Nick's posting:
1. -levels- on SSC picks up the levels of a categorical
variable.
levels state, local(state)
foreach s of local state {
test _Ist`s'XpoXtt
...
}
2. There is an FAQ on this general problem
at http://www.stata.com/support/faqs/data/for.html.
It is showing its age -- it is phrased in terms of -for-
and the -vallist- mentioned there is superseded by
-levels- -- but the discussion may be of interest.
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/
