Skip to content

aggregate.formula: formula from string

4 messages · Thaler,Thorn,LAUSANNE,Applied Mathematics, arun

#
Dear all,

I want to use aggregate.formula to conveniently summarize a data.frame. I have quiet some variables in the data.frame and thus I don't want to write all these names by hand, but instead create them on the fly. This approach has the advantage that if there will be even more columns in the data.frame I don't have to change the code.

I've hence tried to construct a formula object and to pass that to aggregate:

d <- expand.grid(a = factor(1:3), b = factor(LETTERS[1:2]))
d <- rbind(d,d,d)
d$y <- rnorm(18)
d$z <- rnorm(18)
mF <- as.formula(paste("cbind(", paste(names(d)[-(1:2)], collapse = ","), ") ~ a + b", sep = ""))

But if I try to pass that formula to aggregate

aggregate(mF, d, mean)

I get the following error:

Error in m[[2L]][[2L]] : object of type 'symbol' is not subsettable

But if I pass the formula directly:

aggregate(cbind(y, z) ~ a + b, d, mean)

Everything is working as expected.

So I was wondering what went wrong? I know I could use a formula like . ~ a + b instead and this would work fine, but I'm just interested in why the outlined approach does not work as expected, and where my mistake lies? (that means in particular I am not asking for a solution of how to get the thing done - there are plenty of alternatives - but instead to understand why this very approach does not work)

Thanks for your help!

Kind Regards,

Thorn Thaler
Mathematician

Applied Mathematics 
Nestec Ltd,
Nestl? Research Center
PO Box 44 
CH-1000 Lausanne 26
Phone: +41 21 785 8220
Fax: +41 21 785 9486
#
Hi,
Try this:
res<- aggregate(eval(mF),d,mean)

res

#? a b????????? NA????????? NA
#1 1 A -1.48354978 -0.37141485
#2 2 A -0.08862713? 0.35359250
#3 3 A? 1.17519518 -0.47595290
#4 1 B? 0.10214686 -0.70005131
#5 2 B? 0.41185154? 0.03707291
#6 3 B? 0.20507062 -0.67946389


res1<-aggregate(cbind(y, z) ~ a + b, d, mean)
colnames(res)[3:4]<-colnames(res1)[3:4]
?identical(res,res1)
#[1] TRUE
A.K.




----- Original Message -----
From: "Thaler,Thorn,LAUSANNE,Applied Mathematics" <Thorn.Thaler at rdls.nestle.com>
To: "r-help at r-project.org" <r-help at r-project.org>
Cc: 
Sent: Wednesday, October 31, 2012 5:46 AM
Subject: [R] aggregate.formula: formula from string

Dear all,

I want to use aggregate.formula to conveniently summarize a data.frame. I have quiet some variables in the data.frame and thus I don't want to write all these names by hand, but instead create them on the fly. This approach has the advantage that if there will be even more columns in the data.frame I don't have to change the code.

I've hence tried to construct a formula object and to pass that to aggregate:

d <- expand.grid(a = factor(1:3), b = factor(LETTERS[1:2]))
d <- rbind(d,d,d)
d$y <- rnorm(18)
d$z <- rnorm(18)
mF <- as.formula(paste("cbind(", paste(names(d)[-(1:2)], collapse = ","), ") ~ a + b", sep = ""))

But if I try to pass that formula to aggregate

aggregate(mF, d, mean)

I get the following error:

Error in m[[2L]][[2L]] : object of type 'symbol' is not subsettable

But if I pass the formula directly:

aggregate(cbind(y, z) ~ a + b, d, mean)

Everything is working as expected.

So I was wondering what went wrong? I know I could use a formula like . ~ a + b instead and this would work fine, but I'm just interested in why the outlined approach does not work as expected, and where my mistake lies? (that means in particular I am not asking for a solution of how to get the thing done - there are plenty of alternatives - but instead to understand why this very approach does not work)

Thanks for your help!

Kind Regards,

Thorn Thaler
Mathematician

Applied Mathematics 
Nestec Ltd,
Nestl? Research Center
PO Box 44 
CH-1000 Lausanne 26
Phone: +41 21 785 8220
Fax: +41 21 785 9486

______________________________________________
R-help at 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.
#
Dear Arun,

Thanks for your suggestion, that does the trick. Just because I'm curious, where does the problem come from? I figured that wrapping the formula object in brackets would work as well:

aggregate((mF), d, mean)

So I guess it has something to do with the scope of mF, or what is the root cause?

Thanks for your help!


KR,

-Thorn
#
Hi Thorn,

May be it is a bug in aggregate.formula().? Not sure about it.

The solutions that I gave and yours were returning the colnames as NA.? 

I guess this should be much better in terms of getting the result in one step.

?do.call(aggregate,list(mF,d,mean))
#? a b?????????? y?????????? z
#1 1 A -1.48354978 -0.37141485
#2 2 A -0.08862713? 0.35359250
#3 3 A? 1.17519518 -0.47595290
#4 1 B? 0.10214686 -0.70005131
#5 2 B? 0.41185154? 0.03707291
#6 3 B? 0.20507062 -0.67946389
A.K.




----- Original Message -----
From: "Thaler,Thorn,LAUSANNE,Applied Mathematics" <Thorn.Thaler at rdls.nestle.com>
To: arun <smartpink111 at yahoo.com>
Cc: R help <r-help at r-project.org>
Sent: Wednesday, October 31, 2012 8:30 AM
Subject: RE: [R] aggregate.formula: formula from string

Dear Arun,

Thanks for your suggestion, that does the trick. Just because I'm curious, where does the problem come from? I figured that wrapping the formula object in brackets would work as well:

aggregate((mF), d, mean)

So I guess it has something to do with the scope of mF, or what is the root cause?

Thanks for your help!


KR,

-Thorn