Append a vector as a row in a CSV file

Question

What I already have:

Consider the following data — just dummy data, the real data is created automatically:

features <- numeric(0)
features <- c(features, 1, 2, 3, 4, 5, 6)
features2 <- numeric(0)
features2 <- c(features2, 1, 2, 3, 4, 5, 6)
featureVectors <- list()
featureVectors[["file1"]] <- features
featureVectors[["file2"]] <- features2
resultMatrix <- do.call(rbind, featureVectors)
colnames(resultMatrix) <- c("SPEC_MEAN", "SPEC_SD", "SPEC_MODE", "AUTOC_MEAN", "AUTOC_SD", "ZC_MEAN")

This produces an output like the following:

      SPEC_MEAN SPEC_SD SPEC_MODE AUTOC_MEAN AUTOC_SD ZC_MEAN
file1         1       2         3          4        5       6
file2         1       2         3          4        5       6

I can write this to a CSV file by calling write.csv(resultMatrix, file="out.csv").

---

What I need:

As the result file is (currently) just created on the fly, I'd like to write these features (i.e. the vectors) to the CSV file as soon they are evaluated.

So I thought I'd use write.csv and append, but the option isn't available for the method. I then thought I could use write.table, but there are two problems:

1. write.table does not indent the first empty column name, thus leaving my first row shifted one to the left.

2. The data is somehow incorrectly transposed. See example below.

---

What I've tried:

Also, calling these commands …

write.table(resultMatrix, file="data-appended.csv", sep=",")
write.table(features, file="data-appended.csv", sep=",", append=TRUE, col.names=FALSE)

… produces this result:

"SPEC_MEAN","SPEC_SD","SPEC_MODE","AUTOC_MEAN","AUTOC_SD","ZC_MEAN"
"file1",1,2,3,4,5,6
"file2",1,2,3,4,5,6
"1",1
"2",2
"3",3
"4",4
"5",5
"6",6

… which is not what I want. So, how do I append the content of the features vector, including the file name, to an already existing CSV file?

Occasionalism

What's the type of the resultMatrix as R sees it? – Riella

typeof(resultMatrix) => [1] "double". Does it matter, here? I'd just like to write the new feature vectors to the file, i.e. doesn't matter if there's the resultMatrix variable. – Occasionalism

Sorry I was actually looking for output of class(resultMatrix). :) – Riella

Answer 1

Two adjustments will solve your two problems.

• To keep indentation of the column headers, use the unintuitive (but documented!) argument col.names=NA:

  write.table(resultMatrix, file = "data-appended.csv", sep = ",", 
              col.names = NA)
  

• To write features (a vector) as a row rather than a column, transpose it and convert it to a matrix before passing it to write.table():

  FF <- as.matrix(t(features))
  write.table(FF, file = "data-appended.csv", sep = ",", 
              col.names = FALSE, append=TRUE)
  

Babbling

So, is there any easier method to setting the file than: rownames(FF) = fileName to get file1, 1, 2, 3, 4, 5, 6? – Occasionalism

Not really. I think that's the proper way to do this. In response to the earlier incarnation of your comment, you could also use something like the following to create the first header line of the file: writeLines(c("", colnames(resultMatrix)), con = "data-appended.csv", sep = ","). – Nonflammable