R read.table <-> write.table

Question

In R:

d=read.table(filename, as.is = TRUE, header = TRUE, sep = "\t", row.names = 1)

What is the command to write back the exact same file from d?

write.table(d, ?)

I give you one example input file:

one two
1   2

The separator is "\\t". What are the write.table parameters that would write the exact same output file after reading it with read.table?

Thank you, Gregor

Answer 1

The problem is that your read.table used column 1 as row.names so it lost its column name ("one"). When you write it out, you have to do something special to get the "one" name back.

cbind(one=row.names(d), d) will add the row.names as a column with the name "one". Then you simply have to disable the use of row.names and quotes, and specify the separator:

# Create the test file
filename <- "test.txt"
filename2 <- "test2.txt"
cat("one\ttwo\n1\t2\n", file=filename)

# read it in
d <- read.table(filename, as.is = TRUE, header = TRUE, sep = "\t", row.names = 1)

# write it out again
write.table(cbind(one=row.names(d), d), filename2, row.names=FALSE, sep="\t", quote=FALSE)

# Ensure they are the same:
identical(readLines(filename), readLines(filename2)) # TRUE

readLines(filename)
readLines(filename2)

UPDATE To avoid hard-coding the first column name, you mustn't lose it when loading:

# Read the data without any row.names
d <- read.table(filename, as.is = TRUE, header = TRUE, sep = "\t", row.names = NULL)
# Then use the first column as row.names (but keeping the first column!)
row.names(d) <- d[[1]]
d
#  one two
#1   1   2        

# Now you can simply write it out...
write.table(d, filename2, row.names=FALSE, sep="\t", quote=FALSE)

# Ensure they are the same:
identical(readLines(filename), readLines(filename2)) # TRUE

You could of course still remove column 1 if you keep the name of it around and use it as in the first example.

Answer 2

write.table(d,"filename")

you have to specify which file extension you want to use. Hope it works.