Replace every string in dataframe containing character(s)

Question

In the test dataframe below, I am attempting to change every string in the dataframe containing "NA" to "" (so as to make NAs blank).

dat <- as.data.frame(matrix(ncol=2, nrow=2))
dat$V1 <- c("  NA", "foo")
dat$V2 <- c("bar", "NA   ")

dat
   V1   V2
1  NA  bar
2 foo NA

However, the following command returns a completely blank dataframe, as if all strings contained "NA". Why does this happen and what would be the correct solution?

value <- "NA"

dat[grepl(value, dat)] <- ""

To explain Damiano's comment further, you're replacing columns with "" if they contain "NA". Since every column contains "NA", you are of course left with a blank dataframe. Note the difference if you construct a dataframe with dat$V1 <- c(" NA", "NA ") and dat$V2 <- c("foo", "bar ") — aku
– aku, Commented Aug 24, 2017 at 18:10

Damiano Fantini · Accepted Answer · 2017-08-24 18:06:47Z

1

Just using gsub

value <- "NA" 

for (i in 1:ncol(dat)) {
  dat[,i] <- gsub(value, "", dat[,i])  
}
dat

answered Aug 24, 2017 at 18:06

Damiano Fantini

1,97511 silver badges11 bronze badges

Sign up to request clarification or add additional context in comments.

Comments

aku · Accepted Answer · 2017-08-24 18:06:07Z

1

dat <- lapply(dat, function(x) {gsub("NA", "", x)})
dat <- data.frame(dat)

answered Aug 24, 2017 at 18:06

aku

4654 silver badges11 bronze badges

Comments

Eric Watt · Accepted Answer · 2017-08-24 18:05:27Z

0

library(data.table)
setDT(dat)

for(j in seq_along(dat)){
  set(dat, i = which(dat[[j]] %like% "NA"), j = j, value = "")
}
      V1  V2
# 1:     bar
# 2: foo

answered Aug 24, 2017 at 18:05

Eric Watt

3,25012 silver badges21 bronze badges

Comments

MarkusN · Accepted Answer · 2017-08-24 19:43:53Z

0

Maybe in your case you are better off with a matrix.

datm <- as.matrix(dat)

Now your proposed solution works:

datm[grepl(value, datm)] <- ""

or using gsub:

datm = gsub("\\s*NA\\s*", "",datm)

You can convert it to a dataframe after data cleansing.

answered Aug 24, 2017 at 19:43

MarkusN

3,2631 gold badge21 silver badges26 bronze badges

Collectives™ on Stack Overflow

Replace every string in dataframe containing character(s)

4 Answers 4

Comments

Comments

Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

4 Answers 4

Comments

Comments

Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related