<>
So we could get the same effect via
*************
clear*
set obs 1
gen myvar=1
la def yesno 0 "no" 1 "yes"
label values myvar yesno
ta my
decode my, gen(new)
la drop yesno
decode my, gen(new2)
*************
where the problem is now more obvious...
HTH
Martin
-----Ursprüngliche Nachricht-----
Von: [email protected]
[mailto:[email protected]] Im Auftrag von Martin Weiss
Gesendet: Donnerstag, 29. Oktober 2009 15:05
An: [email protected]
Betreff: st: AW: RE: AW: label problem in sts test
<>
Ok, point taken. The quote is still there in [D], p. 355 at the top.
-sts test- tries to -decode- somewhere along the way, and that is where it
chokes on the missing value label...
HTH
Martin
-----Ursprüngliche Nachricht-----
Von: [email protected]
[mailto:[email protected]] Im Auftrag von Isabelle
Deltour
Gesendet: Donnerstag, 29. Oktober 2009 14:57
An: [email protected]
Betreff: st: RE: AW: label problem in sts test
Hi Martin,
Thanks for your answer.
What surprised me is that the documentation says:
"as stated earlier, Stata does not mind if a nonexistent value label is
associated with a variable. When Stata uses such a variable, it simply acts
as if it is not labeled;"
Data management (stata9), p293.
If I am understanding the sentence above correctly, the behavior discussed
should not happen.(and one solution is to label the variable properly).
Hope this helps
Isabelle
-----Original Message-----
From: [email protected]
[mailto:[email protected]] On Behalf Of Martin Weiss
Sent: Thursday, October 29, 2009 2:14 PM
To: [email protected]
Subject: st: AW: label problem in sts test
<>
This problem can be reproduced via
*************
webuse stan3, clear
stset, noshow
sts test posttran
codebook postt
la def yesno 0 "no" 1 "yes"
label values posttran yesno
la drop yesno
sts test posttran
*************
Question: What is surprising about that? Depriving a variable of its -label-
is never a good idea, I would think...
And Isabel already has a remedy available to her, as she mentioned
herself...
HTH
Martin
-----Ursprüngliche Nachricht-----
Von: [email protected]
[mailto:[email protected]] Im Auftrag von Isabelle
Deltour
Gesendet: Donnerstag, 29. Oktober 2009 13:57
An: [email protected]
Betreff: st: label problem in sts test
Dear Stata-listers,
For a survival analysis, I made a dataset where I had assigned a label
To a variable, but, by mistake, the label was not defined.
I think that this situation produced an error message in the
command sts test.
sts test DV
failure _d: diaggrp
analysis time _t: (DateOfFup-origin)/365.25
origin: time fdato
exit on or before: diaggrp==10 20 30 40 50 60 70 80 90 100 110 120
130 140 time mdy(12, 1, 2006)
id: pnr
value label YesNoLbl not found
r(111);
. label list
Is this expected? Have others met this situation?
(Note that I fixed my problem by assigning the label as it should have
been in the first place. )
Thank you for your comments
Isabelle Deltour
*
* 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/
*
* 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/
*
* 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/
