Scanning a csv file for a string in R - r

Question

I would like to be able to scan a csv file row by row in R and exclude the rows that contain the word "target".
The problem is that the data comes from different places and the word "target" can come up in a number of different columns in the data frame.
So I need a line in a function that will look for this string, and if it is not present, then append that row to a new data frame (that I will then write out as a new csv).
Any and all help gratefully recieved.

Answer 1

Andrie's comment is probably the way most users would approach this, but if you want to do this at the reading in stage, you can try this:
Read in your csv using readLines and make any lines that have the text target blank:
temp = gsub(".*target.*", "", readLines("test.csv"))
Use read.table to convert temp to a data.frame. Since all lines that have the text target are now blank, the default blank.lines.skip=TRUE in read.table should correctly read in the rest of your data as a data.frame.
read.table(text=temp, sep=",", header=TRUE)

Answer 2

Use readLines:
lines <- readLines(file)
n.lines <- length(lines)
vec.1 <- rep(0, n.lines)
vec.2 <- rep(0, n.lines)
# more vectors as necessary
counter <- 0
for (i in 1:n.lines){
this.line <- strplit(lines[i], ",")
if ("target" %in% this.line) next
counter <- counter + 1
vec.1[counter] <- this.line[1]
vec.2[counter] <- this.line[2]
# etc.
}
df <- data.frame(vec.1[1:counter], vec.2[1:counter])
You may have to change n.lines slightly and change the indexing of the for loop if your file has headers; two lines would change as follows:
n.lines <- length(lines) - 1
and
for(i in 2:(n.lines+1)){

Answer 3

I would call from.readLines <- readLines(filename) and then just sub-select the rows that don't contain the target string: data <- read.csv(text = from.readLines[-grep('target', from.readLines)], header = F).
The faster way to do it (if your file is huge) would be to grep -v 'target' original.csv > new.csv first on the command line and then run read.csv(new.csv, ...) in R.
But anyway,
> #Without header
> from.readLines <- c('afaf,afasf,target', 'afaf,target,afasf', 'dagdg,asgst,sagga', 'dagdg,dg,sfafgsgg')
> data <- read.csv(text = from.readLines[-grep('target', from.readLines)], header = F)
> print(data)
V1 V2 V3
1 dagdg asgst sagga
2 dagdg dg sfafgsgg
>
> #With header
> from.readLines <- c('var1,var2,var3', 'afaf,afasf,target', 'afaf,target,afasf', 'dagdg,asgst,sagga', 'dagdg,dg,sfafgsgg')
> data <- read.csv(text = from.readLines[-(grep('target', from.readLines[-1]) + 1)])
> print(data)
var1 var2 var3
1 dagdg asgst sagga
2 dagdg dg sfafgsgg