read.csv parameter stringsAsFactors not working?

Question

Based on the documentation read.csv, the parameter stringsAsFactors, when set, should cause quoted data values to be interpreted as factors. Consider the following data file, which we will call test.csv.

"a",b,c 
"1",2,3 
"3",2,3 

When I try to read this data using read.csv, it does not appear to parse the first column as a factor.

foo = read.csv("test.csv", stringsAsFactor=T)
is.factor(foo$a)

Output:

[1] FALSE

I tried to use the column name without quotes, but that did not work either. How can I correct this?

Answer 1

Your example data are coercible to numeric. Try with data that are not so coercible:

foo <- read.csv(text='"a",b,c 
"1",2,3 
"3",2,3 
"a",2,3 ', stringsAsFactors=TRUE)
> foo$a
# [1] 1 3 a
# Levels: 1 3 a

Otherwise use colClasses:

foo <- read.csv(text='"a",b,c 
"1",2,3 
"3",2,3 ', colClasses=c('factor','numeric','numeric'))
> foo$a
# [1] 1 3
# Levels: 1 3

Or you could convert using as.factor after reading the data in.