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

Re: st: Re: Extract variable name in order to change it, iteratively


From   n j cox <n.j.cox@durham.ac.uk>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: Re: Extract variable name in order to change it, iteratively
Date   Wed, 03 Aug 2005 12:23:00 +0100

Jeppe Warberg Larsen asked (minor editing here)

> I have a wide data structure. Each observation consists of a unique
> identifier (say, -id-) and some series of annual variables with a
> two-digit suffix (running in different intervals). Like this:
>
> id var1_88 var1_89 ... var1_02 var2_92 var2_93 ... var2_02
> var3_...................................

> Now I want to -reshape- it to long, but Stata fails to read the
> two-digit values after the turn of century (00 01 02), so I need to
> -rename- all varibles to a four-digit year.

and then later admitted

> the variable names are not actually var1_ var2_ var3_ ... but words:
> something_ somethingelse_

Jeppe got various good advice from Rafa De Hoyos, David Harrison, Philip Ryan and James Muller. Leaving on one side a direct -reshape- solution,
interesting in its own right, but not the main issue here,
the emphasis was on how to set up a -foreach- loop. Something like
this is indicated (this code owes most to David Harrison's solution,
but replaces an -if-/-else- construct by a call to -cond()-):

foreach v of varlist *_?? {
local yy = substr("`v'",-2,2)
local cc = cond(`yy' < 50, 20, 19)
local new = substr("`v'",1,length("`v'")-2) + "`cc'`yy'"
rename `v' `new'
}

I was also interested to see whether -renvars- from STB-61, which purports to offer canned solutions to many renaming problems, was
up to the job.

There may be a simpler way, but a somewhat grotesque one-line solution is possible with -renvars-, here split for display.

renvars *_??, map(substr("@",1,length("@")-2) + cond(substr("@",-2,2) < "50", "20" + substr("@",-2,2),
"19" + substr("@",-2,2)))

A detail of note is the use of -cond()- in both solutions. In the
first case, it is optional but replaces up to 6 lines with one.
In the second case, it is essential as no -if-/-else- construct
can be put inside the -map()- option of -renvars-.

The next issue of the Stata Journal will carry a tutorial on -cond()-
by David Kantor and myself, as it often seems to be overlooked
in problems like these.

Personally, I prefer the -foreach- route here. Brevity is good,
but clarity is more important. I find it easier to follow the
-foreach- logic step by step than to decode a complicated all-in-one solution.

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