Re: st: How to modify the contents of the [if] qualifier in ado files?

[Maarten Buis <maartenlbuis@gmail.com>]

I would not call the program test, as there is already an official
command test in Stata, and having such a command around on your system
can break the official command.

I would not declare the frc option a string, but rather a integer.

I would not use the loop to remove the unwanted observations but rather:

sort `sample', stable
replace `sample' = 0 in 1/`=_N-`frac'-1'

This is probably faster and also solves your problem.

You can preserve the sort order in your data (and that is a good idea)
by adding the option -sortpreserve- to your -program-, see -help
program-.

Hope this helps,
Maarten

2011/12/7  <Tomas.Konecny@cnb.cz>:
> Hi,
> I would like to write a short program that would reestimate a given regression with a smaller number of observations, where
> the number of observations to be removed can be determined by a `frc' option.
>
> For now it looks like this:
>
> program test, nclass
>        version 9.2
>        syntax anything [if] [in], frc(string)
>
>        reg `anything' `if' `in'
>
>        *mark the sample from previous regression
>        tempname sample
>        g `sample'=1 if e(sample)
>
>        *mark observations from the end of the e(sample) that I want to exclude from the next estimation
>        replace `sample'=0 if `sample'[_n+1]==. & `sample'[_n-1]==1 & "`frc'"!=""
>        local i=2
>        while `i'<=`frc' {
>                replace `sample'=0 if `sample'[_n+1]==0
>                local i=`i'+1
>                }
>
>        *NOW HERE IS THE ACTUAL PROBLEM, I would like to have the `sample' selection criteria compatible with other possible [if] constraints,
>        *that is, I would like the program to run also cases like
>
>        if "`if'"!="" {
>        local `if' "`if' & `sample'==0"
>        }
>
>        reg `anything' `if' `in'
>
> end
>
> For example ->
>
> reg pd unempl if time>tq(1994q1), frc(2)
>
> The if "`if'"!="" {.... part onwards, however, does not work.
>
> Thank you very much for any advice,
>
> Tomas
> Czech National Bank
> *
> *   For searches and help try:
> *   http://www.stata.com/help.cgi?search
> *   http://www.stata.com/support/statalist/faq
> *   http://www.ats.ucla.edu/stat/stata/



-- 
--------------------------
Maarten L. Buis
Institut fuer Soziologie
Universitaet Tuebingen
Wilhelmstrasse 36
72074 Tuebingen
Germany


http://www.maartenbuis.nl
--------------------------

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/