Re: st: More elegant code . . .?

[Stas Kolenikov <skolenik@gmail.com>]

On Sat, Jun 4, 2011 at 6:18 PM, Michael Costello
<michaelavcostello@gmail.com> wrote:
> Statalisters,
>
> I have some code that I would like to make more elegant (shorter,
> while still fitting on one line):
> recode letter_auto_stop 0=1 if letter1==0 & letter2==0 & letter3==0 &
> letter4==0 & letter5==0 & letter6==0 & letter7==0 & letter8==0 &
> letter9==0 & letter10==0
>
> The line above works, but it's bulky.  I'd like to find something
> better to use.  I tried:
> recode letter_auto_stop 0=1 if letter1-letter10==0
>
> but stata treated it as "recode letter_auto_stop 0=1 if letter1==0 &
> letter10==0" ignoring letter2 through letter9 entirely.

No. Stata treated it exactly the way you asked: letter1 - letter10 ==
0, meaning letter1==letter10, of course. This is not a varlist, this
is a subtraction.

Here's an option you could consider:

egen byte torecode = anycount( letter1-letter10 ), values(0)
recode  letter_auto_stop 0=1 if torecode

Generally, you should not trust the variable ranges, though, as that
is a way to disaster when for some reason the order of the variables
in the data set may change. However, treating this hidden problem
properly would require a cycle, and with a cycle, you could do things
entirely differently, anyway.

local iflist 1
gen byte torecode = 1
forvalues k=1/10 {
   local iflist `iflist' & letter`k' == 0
   replace torecode = torecode & ( letter`k' == 0 )
}
* either
recode [whatever] if torecode
* or
recode [whatever] if `iflist'

I am not sure either construct qualifies as less bulky, but I would
view either version as more reliable and less error-prone to copying
and pasting.

-- 
Stas Kolenikov, also found at http://stas.kolenikov.name
Small print: I use this email account for mailing lists only.

*
*   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/