write.table writes unwanted leading empty column to header when has rownames

Question

check this example:

> a = matrix(1:9, nrow = 3, ncol = 3, dimnames = list(LETTERS[1:3], LETTERS[1:3]))
> a
  A B C
A 1 4 7
B 2 5 8
C 3 6 9

the table displays correctly. There are two different ways of writing it to file...

write.csv(a, 'a.csv') which gives as expected:

"","A","B","C"
"A",1,4,7
"B",2,5,8
"C",3,6,9

and write.table(a, 'a.txt') which screws up

"A" "B" "C"
"A" 1 4 7
"B" 2 5 8
"C" 3 6 9

indeed, an empty tab is missing.... which is a pain in the butt for downstream things. Is this a bug or a feature? Is there a workaround? (other than write.table(cbind(rownames(a), a), 'a.txt', row.names=FALSE)

Cheers, yannick

Codd

Answer 1

Citing ?write.table, section CSV files:

By default there is no column name for a column of row names. If col.names = NA and row.names = TRUE a blank column name is added, which is the convention used for CSV files to be read by spreadsheets.

So you must do

write.table(a, 'a.txt', col.names=NA)

and you get

"" "A" "B" "C"
"A" 1 4 7
"B" 2 5 8
"C" 3 6 9

Hygienics

@Marek, would it be possible to add a name to the rownames column? I.e., instead of "", add "ID" or something alike? – Lisle

@Lisle From what I know you can't. You may bind row names with data and give them names (as like in question). – Hygienics

@rusalkaguy Your edit has no point. This "extension" is in original question ("workaround other than") – Hygienics

How would you get the number in each column to line up under the col names? – Pounce

@Pounce You mean fixed width format? Look at write.fwf from gdata package. And ask new question instead of comment. And why you ever need that?! – Hygienics

@Hygienics The write.fwf packages doesn't work. The column names are spaced differently from the column values. I need this because I'm trying to write a GAMS file using R and TABLE columns need to overlap their header. – Pounce

@Hygienics you can use tibble to convert rownames to column with a desired ID: tibble::rownames_to_column(var = "ID") – Drastic

@ÖmerAn Well, tibble was not avaiable at the time when answer was posted ;) – Hygienics

Answer 2

A slight modification to @Marek very helpful answer WILL add a header to the rownames column: temporarily add the rownames as the first column in the data.frame, and write that, ignoring the real rownames.

> a = matrix(1:9, nrow = 3, ncol = 3, dimnames = list(LETTERS[1:3], LETTERS[1:3]))
> write.table(data.frame("H"=rownames(a),a),"a.txt", row.names=FALSE)

and you get

"H" "A" "B" "C"
"A" 1 4 7
"B" 2 5 8
"C" 3 6 9

Monanthous

You should edit Marek's answer to include this, I think. – Sneaking

Answer 3

For anyone working in the tidyverse (dplyr, etc.), the rownames_to_column() function from the tibble package can be used to easily convert row.names to a column, e.g.:

library('tibble')
a = as.data.frame(matrix(1:9, nrow=3, ncol=3, 
                  dimnames=list(LETTERS[1:3], LETTERS[1:3])))

a %>% rownames_to_column('my_id')

  my_id A B C
1     A 1 4 7
2     B 2 5 8
3     C 3 6 9

Combining this with the row.names=FALSE option in write.table() results in output with header names for all columns.

Butterball

Answer 4

I revised a simple function from @mnel, which adds flexibility by using connections. Here is the function:

my.write <- function(x, file, header, f = write.csv, ...){
# create and open the file connection
datafile <- file(file, open = 'wt')
# close on exit 
on.exit(close(datafile))
# if a header is defined, write it to the file (@CarlWitthoft's suggestion)
if(!missing(header)) {
writeLines(header,con=datafile, sep='\t')
writeLines('', con=datafile, sep='\n')
}
# write the file using the defined function and required addition arguments  
f(x, datafile,...)
}

You can specify the function to be 'write.table', 'write.csv', 'write.delim' etc.

Cheers!

Dour

Answer 5

I wanted to add an ad-hoc solution that I haven't seen in this question or any linked questions already. Users may want to write a table with the formatting described above, but have a label present at the top of the first column (where the "" is in the accepted answer). The second answer presents a solution which works, but doesn't allow certain characters (and will replace " " characters with "."s, for example). This solution achieves same, but allows any character type permitted by R to be present in the header.

Starting with a matrix 'a' which has the headers you'd like to use, less the header for the row_names column:

headers <- c("row_names", colnames(a)) #create headers, a character vector that will serve as the header line for your output file, "row_names" here can be replaced with a heading of your choice.
test <- data.frame("row_names"=rownames(a),a) #create data frame with "row_names" for top left position, and the existing row names as the first column. 
test <- as.matrix(rbind(headers, test)) #combine prepared headers with data frame above
write(t(test), "test_output.txt", sep = "\t", ncolumns = dim(test)[[2]]) #write output as a tab-separated file. object must be transposed for this command to produce a table in the expected orientation.

Arsis