Re: st: more efficient techniques

[Jeph Herrin <junk@spandrel.net>]

I think any more general solution (Nick has pointed at the best
particular solution here) would require naming your variables
in a consistent and iterative manner. For instance, if you
suffixed your valm* and m* variables appropriately, you could
code:

 foreach s in numlist 1 2 3 4 {
    local i=1
    foreach x of numlist 1 2 12 52 {
       replace hsecostw = valm`s'/`x' if valm`s'>=0 & m`s'==`i'
       local ++i
    }
 }
	
This would not only improve clarity & efficiency, but would likely
be less error prone.

HTH,
Jeph


Michael Alexander wrote:
> Hi Statalist
>
> In the interests of learning some more efficient programming techniques, 
> I am wondering anyone can suggest a way to improve the following code. 
> Ideally, I would like to do away with the second -foreach- statement, 
> which is just a repeat of the first with arguments changed (valm5 for 
> valm3 and m5b for m4). In this case it was no big deal to just copy the 
> first statement and amend its contents, but if I had 20 occurences then 
> I would obviously want to automate the process. I assume a solution 
> would have to do with using an additional -foreach- statement but not 
> quite sure how this would work when I have two arguments that need to 
> change on each loop through.
>
> Thanks in anticipation
>
> gen hsecostw=0 if inlist(m1, 1, 6, 7, 8)
> local i=1
> foreach x of numlist 1 2 12 52 {
>    replace hsecostw = valm3/`x' if valm3>=0 & m4==`i'
>    local ++i
> }
> local i=1
> foreach x of numlist 1 2 12 52 {
>    replace hsecostw = valm5/`x' if valm5>=0 & m5b==`i'
>    local ++i
> }
*
*   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/