Use variable in function to refer to variable in data.frame

Question

I need to apply a series of operations to different variables in a data.frame. To make my code shorter and easier to read, I am trying to functionalize these operations. However, I am having trouble using a variable (in the function) to refer to a variable in the data.frame. Here is a toy example:

set.seed(1)
dframe <- data.frame(ID=rep(1:4, each=3),
                     cond=rep(c("a", "b"), each=6), 
                     x=rnorm(12))
dframe
   ID cond          x
1   1    a -0.6264538
2   1    a  0.1836433
3   1    a -0.8356286
4   2    a  1.5952808
5   2    a  0.3295078
6   2    a -0.8204684
7   3    b  0.4874291
8   3    b  0.7383247
9   3    b  0.5757814
10  4    b -0.3053884
11  4    b  1.5117812
12  4    b  0.3898432

get.means <- function(var, dframe){
  dframe <- dframe[order(dframe$ID),]
  mat    <- aggregate(var~ID, data=dframe, FUN=mean)
  mat 
}
get.means(var=x, dframe=dframe)
Error in eval(expr, envir, enclos) : object 'x' not found

I can get it to work by using get.means(var=dframe$x, dframe=dframe), but this causes other problems down the road. Here are some other things I've tried that haven't worked:

get.means2 <- function(var, dframe){
  dframe <- dframe[order(dframe$ID),]
  mat    <- aggregate(get(var)~ID, data=dframe, FUN=mean)
  mat 
}
get.means2(var=x, dframe=dframe)
Error in get(var) : object 'x' not found

get.means3 <- function(var, dframe){
  dframe <- dframe[order(dframe$ID),]
  mat    <- aggregate(eval(var)~ID, data=dframe, FUN=mean)
  mat 
}
get.means3(var=x, dframe=dframe)
Error in eval(var) : object 'x' not found

get.means4 <- function(var, dframe){
  dframe <- dframe[order(dframe$ID),]
  mat    <- aggregate(dframe[,var]~ID, data=dframe, FUN=mean)
  mat 
}
get.means4(var=x, dframe=dframe)
Error in `[.data.frame`(dframe, , var) : object 'x' not found

zero323 · Accepted Answer · 2013-10-31 00:50:52Z

2

x is defined only in the scope of dframe. You can try this:

get.means <- function(var, dframe){
    aggregate(as.formula(paste(var, '~ ID')), data=dframe, FUN=mean)
}

get.means(var='x', dframe=dframe)

You can also call get.means4 using column name:

get.means4(var='x', dframe=dframe)

or even get.means3 in this way (although I wouldn't recommend that):

get.means3(var=dframe$x, dframe=dframe)

In the last case you safely remove eval and leave:

aggregate(var~ID, data=dframe, FUN=mean)

edited Oct 31, 2013 at 0:50

answered Oct 31, 2013 at 0:30

zero323

331k108 gold badges982 silver badges958 bronze badges

Sign up to request clarification or add additional context in comments.

2 Comments

gung - Reinstate Monica Over a year ago

Thanks for your help. Scope is killing me. I just went through the manual again, & I must have missed something. #3 works w/ my toy example, but breaks down further on w/ the real (ie more complicated) code. The workaround (1) version seems best, but is that how the developers think people should be doing this? It seems awkward.

zero323 Over a year ago

Generally speaking if you use $ syntax it is better to avoid data argument. Although some functions will work correctly with call like this: foobar(df$foo ~ bar, data=df) it can give you unexpected results. In this case use foobar(df$foo ~ df$bar) otherwise df$bar can be interpreted as df$df$bar. Personally I think that using string names is the best solution and should work fine both with dataframes and matrices.

CHP · Accepted Answer · 2013-10-31 01:58:42Z

1

You can consider using data.table instead of data.frame. It allows for smaller and perhaps more manageable code.

require(data.table)
DT <- data.table(ID = rep(1:4, each = 3), cond = rep(c("a", "b"), each = 6), x = rnorm(12))
DT
##     ID cond           x
##  1:  1    a -0.97191681
##  2:  1    a  1.28097125
##  3:  1    a -0.47717701
##  4:  2    a -0.29965951
##  5:  2    a  1.06189839
##  6:  2    a  1.09880286
##  7:  3    b -1.46961507
##  8:  3    b  1.12854103
##  9:  3    b -0.09556682
## 10:  4    b  0.45225307
## 11:  4    b -0.64993127
## 12:  4    b -0.59079915

DT[, mean(x), by = ID]
##    ID          V1
## 1:  1 -0.28531921
## 2:  2  0.01713826
## 3:  3  0.23810745
## 4:  4 -0.16118247

answered Oct 31, 2013 at 1:58

CHP

17.2k4 gold badges42 silver badges59 bronze badges

1 Comment

gung - Reinstate Monica Over a year ago

Thanks. I don't really know data.table & given that this is just a toy example of one small part of more complicated code that I need to fit the solution into, I should really stay with w/ base R. I do intend to try to learn it, though--the vignettes look helpful.

Collectives™ on Stack Overflow

Use variable in function to refer to variable in data.frame

2 Answers 2

2 Comments

1 Comment

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

2 Comments

1 Comment

Your Answer

Sign up or log in

Post as a guest

Related