RE: st: 'Re-ordering' the labels of a variable

["Martin Weiss" <martin.weiss1@gmx.de>]

<>

" The need to make this kind of detail clear is constant!"



My code earlier included the lines 

*******
cap which sencode
if _rc ssc inst sencode
*******

which hopefully limited confusion as to the origin of the command...




" Martin's example achieves that by construction; I saw nothing that made
that certain to be true of Karen's data in the original post."


Very true, I got carried away by the "facility" of my approach via -input-,
so apologies to Ronan Conroy!




HTH
Martin

-----Original Message-----
From: owner-statalist@hsphsun2.harvard.edu
[mailto:owner-statalist@hsphsun2.harvard.edu] On Behalf Of Nick Cox
Sent: Donnerstag, 25. März 2010 19:17
To: statalist@hsphsun2.harvard.edu
Subject: RE: st: 'Re-ordering' the labels of a variable

Let's all be clear on various details here. 

1. -sencode- is user-written and downloadable from various sources, SJ, SSC
and the author's (Roger Newson's) own website. The need to make this kind of
detail clear is constant! 

2. The original purpose of -sencode- was _s_equential encoding in contrast
to -encode-'s default alphabetic encoding. Unless you can be certain that
first occurrences in the data match the desired ordering, then specifying
-label()- _is_ essential for the desired outcome. Martin's example achieves
that by construction; I saw nothing that made that certain to be true of
Karen's data in the original post. 

3. Although -sencode- is undeniably useful, and its existence underlines a
view elsewhere expressed that -encode- should be revisited by StataCorp to
make it more comprehensive, I see nothing here that could not also be
achieved directly with -encode-. 

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

Martin Weiss

No need to define the -label- beforehand:

*******
clear*

cap which sencode
if _rc ssc inst sencode

input str1 stringvar byte freq  
B 3 
A 5
H 5
S 6
Q 4
I 3
end

expand freq
sencode stringvar, gen(numvar)
list, noo

la li numvar
ta numvar

hist numvar, xlab(, valuelabel)
*******

Ronan Conroy

On 25 Márta 2010, at 10:30, Karen Wright wrote:

> I have a diagnosis (string) variable which can take the value: B, A,  
> H, S, Q or I.  Stata (10.0) cleverly changes this order to an  
> alphabetical order, however this does not make sense when trying to  
> display a simple histogram of the distribution.  I would like to  
> force stata to display them in my original order as it is sensible  
> when considering the ordering of the diagnosis catergories.

Use sencode, one of those utilities you simply must have. Define the  
label first, then use it to map the encoding.

. lab def dx 1 "B" 2 "A" 3 "H" 4 "S" 5 "Q" 6 "I",modify
. sencode dx, label(dx) replace



*
*   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/


*
*   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/