Change NA in a factor

Question

So I have data as such.

d <- structure(c(NA, 1L, 1L, 4L, NA, NA, NA, 1L, 4L, 1L), .Label = c("[1,8]", 
"(8,29]", "(29,105]", "(105,738]"), class = "factor")

So I want to change all NA's to 0 and then plot.

library(ggplot2)

ggplot(data = data.frame(var = d)) +
  geom_bar(aes(x = d), fill = "cornflowerblue")

However I try with the following and it doesn't work...

d <- ifelse(is.na(d) == FALSE, "0", d)

Does this answer your question? [Set NA to 0 in R](https://stackoverflow.com/questions/10139284/set-na-to-0-in-r) — Jim G., Jul 30 '20 at 17:47

score 1 · Accepted Answer · edited Jul 30 '20 at 17:07

1

Try this.

levels(d) = append(levels(d), 0)
d[is.na(d)] = 0
ggplot(data = data.frame(var = d)) + geom_bar(aes(x=d), fill="cornflowerblue")

edited Jul 30 '20 at 17:07

Darren Tsai

32,117
5
21
51

answered Jul 30 '20 at 16:46

Haci Duru

456
3
9

excelent! But how would i now put 0 in the front? – John Thomas Jul 30 '20 at 16:55
Thats not exactly correct, that just changed the labels not the data – John Thomas Jul 30 '20 at 17:15
Yes, you are right. I will delete my previous comment. Then I don't know the answer, for now. – Haci Duru Jul 30 '20 at 17:30

dspn · Answer 2 · 2020-07-30T17:07:55.840

there's a couple of issues:

is.na(d) == FALSE is not required, to negate you can use !is.na(d) but since you want to convert NA's you would not want to negate it

secondly you have some issues there because factors are being mixed with other stuff.

you can remove the factors by

d <- levels(d)[d]

then you can remove NAs with

d <- ifelse(is.na(d), "0", d)

finally if you do need the variable to be a factor you can do:

d <- factor(d, ordered = TRUE, levels = c("0", "[1,8]", "(8,29]", "(29,105]", "(105,738]"))

the optional parameter levels was used to prevent some of the levels (that had no records) to be lost

Darren Tsai · Answer 3 · 2020-07-30T17:41:38.283

A base solution with addNA()

d <- `levels<-`(addNA(d), c(levels(d), "0"))

#  [1] 0   [1,8]   [1,8]   (105,738]   0   0   0   [1,8]   (105,738]   [1,8]    
# Levels: [1,8] (8,29] (29,105] (105,738] 0

or using the forcats package

d <- forcats::fct_explicit_na(d, "0")

#  [1] 0   [1,8]   [1,8]   (105,738]   0   0   0   [1,8]   (105,738]   [1,8]    
# Levels: [1,8] (8,29] (29,105] (105,738] 0

Then plot again

ggplot(data = data.frame(var = d)) +
  geom_bar(aes(x = var), fill = "cornflowerblue")

According to your comment, you can put "0" in the front by this way

(forcats::fct_relevel() is more flexible than stats::relevel())

ggplot(data = data.frame(var = relevel(d, "0"))) +
  geom_bar(aes(x = var), fill = "cornflowerblue")

Change NA in a factor

3 Answers3