R-2.9.1, Windows7
Dear list,
I have a question to you that seems very simple to me, but I just can't
figure it out.
I have a dataframe called "ratings" which contains the following
variables: evalR1, evalR2, evalR3, evalR4, scoreR1, scoreR2, scoreR3,
scoreR4, opinionR1, opinionR2, opinionR3, opinionR4. (there are more
variables, but this gives an idea of the data structure).
What I want is run several analyses on all 3 or 4-combinations of a
given variable. So, for example, I want to compute the following ICC's
(function from the psych package):
ICC(cbind(evalR1,evalR2, evalR3))
ICC(cbind(evalR1,evalR2, evalR4))
ICC(cbind(evalR1, evalR3, evalR4))
ICC(cbind(evalR2, evalR3, evalR4))
ICC(cbind(evalR1, evalR2, evalR3, eval4)).
I create a matrix containing the 3-combinations by combn(4,3). Now I
need to call the variables into the function.
First, I tried paste as follows:
combis <- combn(4,3) # this gives the 3-combinations
attach(ratings)
eval <-
paste("evalR",combis[1,1],",evalR",combis[2,1],",evalR","combis[3,1],sep
="")
(this is of course just for 1 combination, as an example)
the output of this is "evalR1,"evalR2,evalR3", but when I run
ICC(cbind(eval)), an error message is given which is not given when I
enter ICC(cbind(evalR1,evalR2, evalR3)) manually. The function appears
not to recognize the variable names. It also does not work to type
ICC(cbind(unquote(eval))).
Alternatively, I have tried the cat function, but also here ICC does not
recognize the input as variable names.
What am I doing wrong? How can I automatically construct the set of
variable names such that a function recognizes them as variable names?
ICC is one example, but there are also other computations to be run and
the set of variables is pretty large, so typing the combinations of
variable names manually is really unattractive.
What am I missing? It seems to me that there probably is a very simple
solution in R, but which?
Thank you,
Peter Verbeet
[[alternative HTML version deleted]]
Hi,
It's not enough to create a string with your instructions, it also needs to
be evaluated as such.
If you really wanted to evaluate your string, you'd need something like,
a <- b <- cc <- 1 # dummy example
eval(parse(text = "cbind(a, b, cc)"))
#library(fortunes)
#fortune("parse")
but fortune-ately, you probably oughtn't do that. Instead, I think you could
use this example,
d = data.frame(a1=1:10, b1=sin(1:10), c1 = cos(1:10)) # some data.frame
variables = paste(letters[1:3],"1",sep="") # its variable
names
(alternatively, names(d) would do)
ICC = function(x) x # your function
ICC(d[ variables[ c(1, 3) ] ]) # apply the function to a subset of the data,
by selecting the required columns
(If I understood correctly your question)
HTH,
baptiste
2009/9/7 Helter Two <heltertwo@care2.com>
> R-2.9.1, Windows7
>
> Dear list,
>
> I have a question to you that seems very simple to me, but I just can't
> figure it out.
> I have a dataframe called "ratings" which contains the following
> variables: evalR1, evalR2, evalR3, evalR4, scoreR1, scoreR2, scoreR3,
> scoreR4, opinionR1, opinionR2, opinionR3, opinionR4. (there are more
> variables, but this gives an idea of the data structure).
>
> What I want is run several analyses on all 3 or 4-combinations of a
> given variable. So, for example, I want to compute the following ICC's
> (function from the psych package):
> ICC(cbind(evalR1,evalR2, evalR3))
> ICC(cbind(evalR1,evalR2, evalR4))
> ICC(cbind(evalR1, evalR3, evalR4))
> ICC(cbind(evalR2, evalR3, evalR4))
> ICC(cbind(evalR1, evalR2, evalR3, eval4)).
>
> I create a matrix containing the 3-combinations by combn(4,3). Now I
> need to call the variables into the function.
> First, I tried paste as follows:
> combis <- combn(4,3) # this gives the 3-combinations
> attach(ratings)
> eval <-
>
paste("evalR",combis[1,1],",evalR",combis[2,1],",evalR","combis[3,1],sep
> ="")
> (this is of course just for 1 combination, as an example)
> the output of this is "evalR1,"evalR2,evalR3", but when I
run
> ICC(cbind(eval)), an error message is given which is not given when I
> enter ICC(cbind(evalR1,evalR2, evalR3)) manually. The function appears
> not to recognize the variable names. It also does not work to type
> ICC(cbind(unquote(eval))).
> Alternatively, I have tried the cat function, but also here ICC does not
> recognize the input as variable names.
>
> What am I doing wrong? How can I automatically construct the set of
> variable names such that a function recognizes them as variable names?
> ICC is one example, but there are also other computations to be run and
> the set of variables is pretty large, so typing the combinations of
> variable names manually is really unattractive.
>
> What am I missing? It seems to me that there probably is a very simple
> solution in R, but which?
>
> Thank you,
> Peter Verbeet
>
>
> [[alternative HTML version deleted]]
>
> ______________________________________________
> R-help@r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
> http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
>
--
_____________________________
Baptiste Auguié
School of Physics
University of Exeter
Stocker Road,
Exeter, Devon,
EX4 4QL, UK
http://newton.ex.ac.uk/research/emag
______________________________
[[alternative HTML version deleted]]
may be this can work
testfun<-function(x) {
rval=""
k<-length(x)
for (i in 1: k) rval<-paste(rval,x[i],sep="-")
rval
}
v1<-paste("evalr",1:4,sep="")
eval<-expand.grid(w=v1,x=v1,y=v1,z=v1)
n<-dim(eval)[1]
results<-rep("", n)
for (i in 1:n) {
row<-unique(unlist(eval[i,]))
if (length(row)>=3) results[i]<-testfun(row)
}
You just have to replace testfun by your own function in this case ICC.
Sincerly.
Justin BEM
BP 1917 Yaoundé
Tél (237) 76043774
________________________________
De : Helter Two <heltertwo@care2.com>
À : r-help@r-project.org
Envoyé le : Lundi, 7 Septembre 2009, 18h17mn 22s
Objet : [R] calling combinations of variable names
R-2.9.1, Windows7
Dear list,
I have a question to you that seems very simple to me, but I just can't
figure it out.
I have a dataframe called "ratings" which contains the following
variables: evalR1, evalR2, evalR3, evalR4, scoreR1, scoreR2, scoreR3,
scoreR4, opinionR1, opinionR2, opinionR3, opinionR4. (there are more
variables, but this gives an idea of the data structure).
What I want is run several analyses on all 3 or 4-combinations of a
given variable. So, for example, I want to compute the following ICC's
(function from the psych package):
ICC(cbind(evalR1,evalR2, evalR3))
ICC(cbind(evalR1,evalR2, evalR4))
ICC(cbind(evalR1, evalR3, evalR4))
ICC(cbind(evalR2, evalR3, evalR4))
ICC(cbind(evalR1, evalR2, evalR3, eval4)).
I create a matrix containing the 3-combinations by combn(4,3). Now I
need to call the variables into the function.
First, I tried paste as follows:
combis <- combn(4,3) # this gives the 3-combinations
attach(ratings)
eval <-
paste("evalR",combis[1,1],",evalR",combis[2,1],",evalR","combis[3,1],sep
="")
(this is of course just for 1 combination, as an example)
the output of this is "evalR1,"evalR2,evalR3", but when I run
ICC(cbind(eval)), an error message is given which is not given when I
enter ICC(cbind(evalR1,evalR2, evalR3)) manually. The function appears
not to recognize the variable names. It also does not work to type
ICC(cbind(unquote(eval))).
Alternatively, I have tried the cat function, but also here ICC does not
recognize the input as variable names.
What am I doing wrong? How can I automatically construct the set of
variable names such that a function recognizes them as variable names?
ICC is one example, but there are also other computations to be run and
the set of variables is pretty large, so typing the combinations of
variable names manually is really unattractive.
What am I missing? It seems to me that there probably is a very simple
solution in R, but which?
Thank you,
Peter Verbeet
[[alternative HTML version deleted]]
______________________________________________
R-help@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
[[alternative HTML version deleted]]