reading file using fread with row and column names

Question

Using fread, how to read CSV file which contains row and column names. I tried following but it is not reading the row and column names properly.

The csv file looks like (where C1,C2,C3 are column names and r1, r2, r3 are row names)

input = ",C1,C2,C3
r1,A,B,C
r2,1,3,5
3,2,4,6"

I use function

require(data.table)
fread(input,header = TRUE)

which give

   r1 A B C
1: r2 1 3 5
2:  3 2 4 6

How can I properly read CSV using fread?

Have you tried adding `row.names = 1` when using `read.csv`? If not, you should take a look at `?read.csv` — Eric Fail, Aug 07 '14 at 14:00
@EricFail - Sorry I am new to R/data.table .. What does it has to do with read.csv? — d.putto, Aug 07 '14 at 14:06
have a look at [this SO tread](http://stackoverflow.com/a/13271846/1305688). — Eric Fail, Aug 07 '14 at 14:07
I want to use fread because my data file is big and I do not want to use read.csv . — d.putto, Aug 07 '14 at 14:10
This seems to be fixed in [v1.9.3](https://github.com/Rdatatable/data.table).. — Arun, Aug 07 '14 at 17:07

score 5 · Answer 1 · answered Aug 07 '14 at 14:20

5

You should submit a bug report.

Here is a work-around:

colnames <- strsplit(readLines(textConnection(input), n=1), ",")[[1]]
colnames[1] <- "rownames"
setnames(DT <- fread(input, skip=1, header=FALSE), colnames)
DT
#   rownames C1 C2 C3
#1:       r1  A  B  C
#2:       r2  1  3  5
#3:        3  2  4  6

As you should know, data.table doesn't support rownames.

answered Aug 07 '14 at 14:20

Roland

117,893
9
163
255

@d.putto, it would be interesting to know if `fread()` is still faster in this case. – Eric Fail Aug 07 '14 at 14:28
1

@EricFail Of course it is for a reasonably big file. You don't seem to know much about `fread` (and data.table). – Roland Aug 07 '14 at 14:32
I don't. I hadn't heard of `fread` until 50 minutes ago. I am here to learn. Thanks. – Eric Fail Aug 07 '14 at 14:35

reading file using fread with row and column names

1 Answers1

Linked