Question

I have a matrix with numeric column labels, for example:

1,2,3
1,2,3
4,5,6
7,8,9

When I load this data into a variable in R using as.matrix(read.table("myfile", sep=",", header=TRUE) and print the resulting matrix, the column headers have been prepended with X, and this extra character appears in plots and the like:

     X1     X2     X3
[1,]  1      2      3
[2,]  4      5      6
[3,]  7      8      9

How can I stop this behaviour?

Answer 1

Supply check.names=FALSE to read.table.

Be aware that when data.frames have non-standard names such as this any reference to their columns must be surrounded with tick marks, e.g.

my.df <- read.table(text='1,2,3
1,2,3
4,5,6
7,8,9', header=TRUE, sep=',', check.names=FALSE)

my.df$`1`
with(my.df, `1`)
attach(k)
`1`

Also, calls to transform, such as transform(my.df,1=1:3), will replace the column names with their standardized forms (X1, X2, X3) because transform calls data.frame() with its default arguments.

Setting check.names=FALSE, while useful in some very particular situations, will allow duplicate column names and non syntactic names, which could cause any number of problems later on. Be cautious when using this.

data.frame() uses check.names=TRUE by default because in many contexts R tries to interpret the column names of a data frame as variables in their own right; if the column names are not syntactically valid names (see ?make.names), or protected with back-ticks as described above, then errors will occur in these contexts.