Question

First, this is a very basic question that I’m unsure of how to phrase. If the question is a duplicate (though I checked using what I thought might be appropriate phrasing), I’ll obviously retract and appreciate the link.

Second, I am sure there is an easier way to do what I’m trying, but don’t want to get off-track.

OK. I’m attempting to just get a table of column proportions from a matrix of 0/1’s (the proportion of 1’s conditional on a value of another variable, which is PARTY in this case).

my data.frame is m103, and of dimensions (437,91) and the following process works (as in, produces what I want):

prop.table(as.matrix(ddply(m103, .(PARTY), sum, na.rm=T)))

But of course, I want to actually keep the output, and this is where the error arises. If I do this:

a <- prop.table(as.matrix(ddply(m103, .(PARTY), sum, na.rm=T)))

Things are great. But IMMEDIATELY after this, if I try:

m103.avg.prop <- prop.table(as.matrix(ddply(m103, .(PARTY), sum, na.rm=T)))

I get the error:

Error in FUN(X[[2L]], ...) : only defined on a data frame with all numeric variables

I’d like to keep a rational naming scheme going in my code (which the second example would continue), but I can’t tell if this has something to do with what I’ve tried to assign the output to, or something else.

Many thanks!

EDIT: Let’s see if I can be more explicit

#Data import
m103 <- read.csv("103_members_party.csv", header=T)
#See the first few rows/columns
m103[1:5,1:5]
#Produces this: 
     ID PARTY X930 X461 X137
1 15245   100    0    0    0
2 15000   100    0    0    0
3 29108   200    0    0    0
4 15001   100    0    0    0
5 29132   100    0    0    0
#Sum and get col percentages by PARTY (sums the 1's when PARTY==100, PARTY==200, etc)
#WITHOUT assigning to anything
prop.table(as.matrix(ddply(m103, .(PARTY), sum, na.rm=T)))
#Produces: 
            PARTY          V1
[1,] 1.122515e-05 0.580000465
[2,] 2.245030e-05 0.416619418
[3,] 3.681849e-05 0.003309623
#With assignment to a
a <- prop.table(as.matrix(ddply(m103, .(PARTY), sum, na.rm=T)))
a
#Produces
            PARTY          V1
[1,] 1.122515e-05 0.580000465
[2,] 2.245030e-05 0.416619418
[3,] 3.681849e-05 0.003309623
#Now, assignment to m103.avg.prop
m103.avg.prop <- prop.table(as.matrix(ddply(m103, .(PARTY), sum, na.rm=T)))
#results in error: 
Error in FUN(X[[2L]], ...) : 
  only defined on a data frame with all numeric variables

Answer 1

The error you’re getting is because you’re trying to sum something that isn’t a number. Without reproducible code I can’t tell you exactly what is going on. But, one of the reasons we ask for a reproducible example is that in the process of making one, you will often discover the problem on your own.

In this case, I assume the data came from somewhere like excel, which is notorious for doing surprising things to data. try looking at str(m103) and one of the column will be a character vector rather than numeric. faulting that, i would have to see your data.

However, there should be no difference between your assignment to a and your assignment to m103.avg.prop. As a side note, I like to avoid numbers in my variable names wherever possible, just to avoid confusing myself!

EDIT: Add runable code:

> m103<-data.frame(ID=c(15245, 15000, 29108, 15001, 29132),PARTY=c(100, 100, 200, 100, 100),X930=c(0, 0, 1, 0, 0),X461=c(0, 0, 0, 1, 1),X137=c(1, 1, 1, 1, 1))
> m103
     ID PARTY X930 X461 X137
1 15245   100    0    0    1
2 15000   100    0    0    1
3 29108   200    1    0    1
4 15001   100    0    1    1
5 29132   100    0    1    1

> prop.table(as.matrix(ddply(m103, .(PARTY), sum, na.rm=T)))
            PARTY        V1
[1,] 0.0009579095 0.7163630
[2,] 0.0019158189 0.2807633
> a <- prop.table(as.matrix(ddply(m103, .(PARTY), sum, na.rm=T)))
> m103.avg.prop <- prop.table(as.matrix(ddply(m103, .(PARTY), sum, na.rm=T)))
> a
            PARTY        V1
[1,] 0.0009579095 0.7163630
[2,] 0.0019158189 0.2807633
> m103.avg.prop
            PARTY        V1
[1,] 0.0009579095 0.7163630
[2,] 0.0019158189 0.2807633
> 

I still cannot replicate your problem. Like I said above, the output of str(m103) and the output of str(a) will be informative. Also, sessionInfo(). Short of that, I’ll stick with my previous guesses…