Re: st: RE: Dealing with string variables

[Ron�n Conroy <rconroy@rcsi.ie>]

Nick Cox wrote:

> drop if grain == "wheat" 
> value labels for string variables: 
> not allowed 
> what you want is better done by 
> gen Crop = "crop" if inlist(crop, "wheat", "barley") 
> replace Crop = "root" if inlist(crop, "potato", "cassava", "yams", "beet") 
> ... 
>  
using -encode-

. label define crop 1 "wheat" 2 "barley" 3 "root", modify
. encode crop, gen(Crop) label(crop)
. replace Crop=3 if inlist(crop, "potato", "cassava", "yams", "beet")
. replace Crop=. if Crop > 3

Logic:
1. start by defining a label that will assign numeric codes to the 
strings in a logical order. In a do-file, always follow this by -modify- 
in case the label is already defined in some way.
2. use -encode- to generate a new variable, using the predefined 
mappings. Strings that are not found in the predefined value label will 
be added to the value label, using unassigned numbers. For this reason,  
nothing will be assigned the number 3 because we reserved this for a new 
category, "root".
3. Nick's nice piece of code that checks for root crops. You could use 
the -index- function for a single string, but -inlist- is more general. 
-index- has some neat uses; this isn't one of them.
4. Any remaining values are ones that we don't need (strings like "not 
applicable" or other crops that don't interest us). These will have been 
assigned value labels and numeric codes that can be inspected easily by 
typing

. lab list crop




--

Ronan M Conroy (rconroy@rcsi.ie) 
Senior Lecturer in Biostatistics 
Royal College of Surgeons 
Dublin 2, Ireland 
+353 1 402 2431 (fax 2764) 
-------------------- 
Just say no to drug reps 
http://www.nofreelunch.org/

*
*   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/