Stata The Stata listserver
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

st: RE: RE: RE: create many variables at once


From   "Nick Cox" <n.j.cox@durham.ac.uk>
To   <statalist@hsphsun2.harvard.edu>
Subject   st: RE: RE: RE: create many variables at once
Date   Tue, 23 Jul 2002 16:46:01 +0100

Joly.Patrick@ic.gc.ca commented on various solutions with -for*-:

> The above solutions reminded me that -for-'s functionality
> hasn't quite been
> subsumed by the introduction of -foreach- and -forvalues-
> as I have been
> prone to believe.  In version 6, I was quite content using
> the set of
> looping commands (tokenize, while, etc.) and hadn't got to
> learning -for-
> just yet.  Thus, when -foreach- and -forvalues- came about
> with Stata 7, I
> saw no reason to bother with it.   But the solutions are excellent
> illustrations that -for-'s sometimes results in more
> succinct and, arguably,
> elegant solutions as it avoids unnecessary looping.
>
> To right the wrong, I looked up for.hlp once more and noted small
> inconsistency the examples provided.  The last example reads:
>
> (begin excerpt)
>         . for any male female: save c:\data\X if sex=="X"
>
> will not work.  for will think there are three commands:
>
>         save c:
>         data
>         X if sex=="X"
>
> Instead, you must type
>
>         . for any male female: save "c:\data\X" if sex=="X"
> (end of excerpt)
>
> The last line won't work either since the syntax of save
> doesn't allow  the
> [if exp] condition -- a non-feature that I, among others,
> wouldn't mind
> seeing implemented in -save- BTW. ;)  (NB: I know it's been
> discussed on the
> list before)
>

I agree that the -for- solution is neater here for the problem posed.

However, much of the engineering work behind -for- was done precisely
to achieve this result.

That is, -for- is optimised for what I think of as "parallel"
problems:

do something with the ith member of this list
[ and the ith member of that list
[ and the ith member of THAT list
...
]
]

for i = 1, ..., I

so that there are I results.

In contrast, -foreach- and -forvalues- grow very well indeed with
"nested" problems:

do something with the ith member of this list
[ and the jth member of that list
[ and the kth member of that list
...
]
]

for i = 1, ..., I
    j = 1, ..., J
    k = 1, ..., K
    ...

so that there are I * J * K * ... results.

On the whole, I think nested problems are more common in
(my) Stata practice.

On the face of it, it is difficult to devise a simple syntax
to cope with both kinds of problems. Some would argue that
it is difficult to devise a simple syntax to cope with parallel
problems.

Having said that, note that Kit Baum provided an
approach using -foreach- to the original problem.

Nick
n.j.cox@durham.ac.uk

*
*   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/



© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index