Converting DF columns to factor is less than straightforward

Question

I'm learning R and would welcome some explanation/pointers as to what the error means and why it is that I can't simply assign the columns as factors:

In R, I read.csv() a file into a variable DF. class(DF) tells me it's a "data.frame" but columns 1, 2, and 3 are non-factors. When I try to assign columns 1,2,3 as factors I get an error:

asFactors <- c(1:3)
DF[asFactors] <- as.factors(DF[asFactors]) # same if I use DF[,asFactors] 
Error in sort.list(y) : 'x' must be atomic for 'sort.list'
Have you called 'sort' on a list?

What does sort have to do with me trying to change variable type?

Instead, what I seem to need to do is use an apply() function to convert columns to factors (but this isn't the most intuitive thing):

DF[,asFactors] <- lapply(DF[asFactors], factor)

Furthermore, if I try to convert columns to 4,5,6 to numeric using lapply i get a new error:

asNumeric <- c(4:6)
DF[,asNumeric] <- lapply(DF[asNumeric], numeric) 
Error in FUN(X[[i]], ...) : invalid 'length' argument

And if I fall back on my original attempt, I get:

DF[,asNumeric] <- as.numeric(DF[,asNumeric])
Error: (list) object cannot be coerced to type 'double'

So for each variable type I seems= to need a different method of converting columns, or I haven't found the one method that applies to all of them.

`lapply` should work. You probably have problems in the data. Provide a reproducible example for people to work with. — Rich Scriven, Sep 11 '17 at 05:21
https://stackoverflow.com/questions/33180058/coerce-multiple-columns-to-factors-at-once — Rich Scriven, Sep 11 '17 at 05:21
lapply does work for factors but not for numeric. I take it you mean there may be issues with the numeric data. — val, Sep 11 '17 at 05:22
@thelatemail: that was it. Why isn't it consistent with "factor" in the first case? factor and numeric or as.factor and as.numeric would make more sense to me. — val, Sep 11 '17 at 05:25
@val - `as.factor` also exists. It just so happens that `factor` is able to generate a factor given your input. — thelatemail, Sep 11 '17 at 05:26
You should study `help("[")`. Your first attempt doesn't work, because it tries to convert a data.frame to factor. — Roland, Sep 11 '17 at 06:04

www · Accepted Answer · 2017-09-11T05:36:22.680

3

To change multiple columns to factor, use:

DF[,1:3] <- lapply(DF[,1:3], factor)

To change from factor to numeric, remember to use as.numeric(as.character(x)), like this:

DF[,1:3] <- lapply(DF[,1:3], function(x) as.numeric(as.character(x)))

edited Sep 11 '17 at 05:36

answered Sep 11 '17 at 05:18

www

4,124
1
11
22

lapply did work for factors but not for numeric. – val Sep 11 '17 at 05:23
@val - See if the above works for converting to numeric now. – www Sep 11 '17 at 05:31

Converting DF columns to factor is less than straightforward

1 Answers1