fwrite in data.table changed my string column to integer, how to fix?

Question

I'm writing a data table to csv using fwrite. The function changes the class of one column from character to integer and deleted the starting zeros, which I do not want. So my var2 here is character, and I would like to preserve that through fwrite then fread a subset.

However, what I got was the following:

head(DT)
var1 var2
 1  "0012"
 2  "0032"
 3  "0043" 

DT1 <- DT[var1!=2,]

fwrite(DT1,"DT1.csv", row.names=FALSE)

Then:

DT1 <- fread("DT1.csv")
head(DT1)
var1 var2
 1    12
 3    43 

Is there any way I can fix it?

Answer 1

The issue is with fread rather than fwrite. So just specify that var2 is a character class. The default in fread is sensing that var2 is a numeric column so just add the following in your call to fread:

DT1<-fread("DT1.csv", colClasses = list(character=c("var2")))

Let me know if this works for you and good luck!