0

Consider a example:

library(data.table)
myDt <- data.table(a=rep(1:10), b=paste(1:10, "000",sep=","), c=paste(2:11, "000", sep=",") )
myDt

Now I want to convert the character columns into numeric. Basically I want to remove the commas. I have already figured a way. 

charToNumCols <- c("b", "c")

numColsDt <- myDt[, lapply(.SD, function(x) as.numeric(gsub(",", "", x)) ), .SDcols=charToNumCols]

newDt <- data.table(myDt[, .SD, .SDcols=setdiff(names(myDt), charToNumCols)],     numColsDt)

Now I want a solution in which I dont have to create an intermediate data.table. Originally for this I wrote the following R code

myDt <- myDt[, charToNumCols:=lapply(.SD, function(x) as.numeric(gsub(",", "", x)) ), .SDcols=charToNumCols]

But this does not do the required operation.

Can anyone answer what should be the most efficient coding for the above, considering data.table in R?

mtoto
  • 23,919
  • 4
  • 58
  • 71
Kumar Manglam
  • 2,780
  • 1
  • 19
  • 28

0 Answers0