[19034] in s-news-athena
=?iso-8859-1?Q?R=E9f._:_[S]_Summary:_Adding_levels_to_a_factor?=
daemon@ATHENA.MIT.EDU (=?iso-8859-1?Q?"G=E9rald_Jean"?=)
Tue Aug 24 14:31:59 1999
From: =?iso-8859-1?Q?"G=E9rald_Jean"?= <Gerald.Jean@spgdag.ca>
To: s-news@wubios.wustl.edu
Message-Id: <852567D7.006430E5.00@mail.spgdag.com>
Date: Tue, 24 Aug 1999 14:30:39 -0400
Mime-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
Thanks to prof. Ripley,
of course his solution works and all the factors can be done at once.
> ttt.new_na.tree.replace(ttt.po2[,index.of.factor.variables])
Thanks again
Gérald Jean
Analyste-conseil (statistiques), Actuariat
télephone : (418) 835-8839
télecopieur : (418) 835-5865
courrier électronique: gerald.jean@spgdag.ca
"In God we trust all others must bring data"
"Gérald Jean" <Gerald.Jean@spgdag.ca> le 99/08/24 13:49:09
Pour s-news@wubios.wustl.edu
:
cc : (ccc : Gérald Jean/Assurances/SPGDAG/Desjardins)
Objet [S] Summary: Adding levels to a factor
:
Thanks to all who responded (Todd Taylor, Keenan Sean, Samuel Buttrey,
Leonid Gibiansky, S.D. Byers, Bill Dunlap, Don McKenzie and Yuelin Li) the
original message follows. All suggested a similar approach, i.e. first
convert the factor to character and then convert back to factor adding the
new level in the process; one thing I forgot to say is that I was not
creating the factor variables myself, there is several with the same
problem, they were created by someone else in a SAS program and I imported
the data set using Frank Harrel's "sas.get" function. Here is an exemple
solution from Todd Taylor:
new <- factor(as.character(old),
levels=c(levels(old), ""),
labels=c(levels(old), "Missing")
)
Again thanks to all,
Gérald Jean
Analyste-conseil (statistiques), Actuariat
télephone : (418) 835-8839
télecopieur : (418) 835-5865
courrier électronique: gerald.jean@spgdag.ca
"In God we trust all others must bring data"
The original message:
Hi there,
Suppose I have a factor variable with 2 levels.
> levels(ttt.dcontra)
[1] "N" "O"
But in fact there is a whole lot of missing values in that variable.
> summary(ttt.dcontra)
N O NA's
5639 7468 4593
The missing values mean something to me and I'd like to assign them some
factor level,
"Missing" say, that S+ will treat as a valid level in further analysis, in
regression
for exemple. How do I do that?
If I try to assign a value to the "NA's", S+ complains.
> ttt.dcontra[which.na(ttt.dcontra)] _ 'Missing'
Warning messages:
replacement values not all in levels(x): NA's generated in:
"[<-.factor"(ttt.d
contra, which.na(ttt.dcontra), value = "Missing")
If I try to assign a new level to the variable, S+ complains again.
> levels(ttt.dcontra) _ edit(levels(ttt.dcontra))
Error in "levels<-.factor"(ttt.dcontra, edit(level..: number of levels
differs:
2 old, 4
I don't have anymore ideas, any help would be appreciated, I'll summarize
to the list.
-----------------------------------------------------------------------
This message was distributed by s-news@wubios.wustl.edu. To unsubscribe
send e-mail to s-news-request@wubios.wustl.edu with the BODY of the
message: unsubscribe s-news
-----------------------------------------------------------------------
This message was distributed by s-news@wubios.wustl.edu. To unsubscribe
send e-mail to s-news-request@wubios.wustl.edu with the BODY of the
message: unsubscribe s-news
