Thanks very much Austin, I have looked into your suggestion: unfortunately
it does not run appreciably faster.
No, I have no idea about how many variables are being generated with -xi-
in each iteration, but it is certainly far more than necessary!
I must explain that the dataset is very large (in the millions) but runs in
a matter of minutes with sas, so I expect that there must be a way to
improve the computational speed.
So I am still very much open to suggestions, if anyone out there has
experience in coding loops involving -xi- with large datasets, I'd love to
hear from you. I'm a stata novice.
By the way there was an unnecessary semi-colon in the code you sent out
Austin (and produces an error message), the code should read:
xi i.age i.educ i.educ*age i.educ*age2 i.educ*age3 i.educ*age4
rather than
xi: i.age i.educ i.educ*age i.educ*age2 i.educ*age3 i.educ*age4
I owe this very useful tip to a statalist exchange from 2004:
http://www.datasets.org/statalist/archive/2004-10/msg00413.html, summed up
here:
**********************
When you use -xi- in combination with another command, for example,
with -regress-, you must use the : (colon character).
xi: regress y i.age
But when you use -xi- alone with just the variable, then you must not use
the colon character.
xi i.age <-- notice that there is no colon in this usage of -xi-
regress y _I*
If you violate the rule, for example,
xi: i.age
then Stata will actually do the dummy variable creation, but it will also
complain about the violation of syntax rules. With the colon character,
Stata expects the immediately following text to be another command, such
as -regress-. The complaint (error message) is that _Iage_ is not the name
of any command that Stata can find.
Joseph Coveney
**********************
And yes, Austin, you are right that I intented to weigh the observations,
but this got deleted in the code I sent out.
Patrick.
"Austin Nichols" <[email protected]>
Do you realize you're generating many variables with -xi- in each
iteration of the loop?
Do you want something like the following?
xi: i.age i.educ i.educ*age i.educ*age2 i.educ*age3 i.educ*age4
forv i=94/102 {
logit yr00 _I* [pw=ihwt] if year~=`i'
predict pyr00 if e(sample), p
gen rw00=ihwt*pyr00/(1-pyr00)
sum yr00 pyr00 rw00 if year~=100
drop pyr00 rw00
}
Also the -sum- command is unweighted--is that what you intended?
On Thu, Apr 3, 2008 at 3:24 AM, <[email protected]> wrote:
> Hi there,
>
> I am trying to replicate a loop that runs well and fast in SAS, but is
very
> slow in STATA on my machine.
>
> Perhaps someone will spot some awkward piece of code in the following?
>
> many thanks for your attention,
>
> [email protected]
>
>
> *** STATA PROGRAM
> set memory 700m ** couldn't do better than 700M
> version 10
> use r2f3_glm.dta, clear
>
> local i 94
> while `i' ~= 103 {
> xi: logit yr00 i.age i.educ i.educ*age i.educ*age2 i.educ*age3
i.educ*age4
> [pweight=ihwt] if year~=`i'
> predict pyr00 if e(sample), p
> gen rw00=ihwt*pyr00/(1-pyr00)
> sum yr00 pyr00 rw00 if year~=100
> drop pyr00 rw00
> local i = `i'+ 1
> }
>
> exit
>
>
> *** SAS PROGRAM
>
> proc logistic data=one(where=(year ne 100)) descending;
> class yr00 educ agedum;
> model yr00=agedum educ educ*age educ*age2 educ*age3 educ*age4;
> weight ihwt;
> output out=temp1 predicted=pyr00;
> run;
>
> data temp1;
> set temp1;
> rw00=ihwt*pyr00/(1-pyr00);
>
> *
> * 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/
>
*
* 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/
*
* 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/
