I have a problem that should be easy to solve but I simply cannot figure it out. I have a huge dataset with groups and a variable. Some groups are empty for this variable (filled only with NAs) and some contains values but also NAs.
For example:
ID <- c("A1","A1","A1","A1","B1","B1","B1","B1", "B1", "C1", "C1", "C1")
Value1 <- c(0,2,1,1,NA,1,1,NA,1,NA,NA,NA)
data <- data.frame(ID, Value1)
I would like to change all NAs to zeros but only in groups that otherwise contain information.
So like this:
ID <- c("A1","A1","A1","A1","B1","B1","B1","B1","B1","C1","C1","C1")
Value1 <- c(0,2,1,1,0,1,1,0,1,NA,NA,NA)
I tried to use group_by(ID) and "replace" with the condition max(Value1)>=0 but either max() doesn't work as a condition or it doesn't work with NAs. Unfortunately I would need this kind of conditioning often in my work so I would also appreciate any suggestions on which are the best packages to treat groups selectively.
You can use a simple if` statement, ie
library(dplyr)
library(tidyr)
data %>%
group_by(ID) %>%
mutate(Value1 = if (all(is.na(Value1))){Value1}else{replace_na(Value1, 0)})
which gives,
# A tibble: 12 x 2 # Groups: ID [3] ID Value1 <fct> <dbl> 1 A1 0 2 A1 2 3 A1 1 4 A1 1 5 B1 0 6 B1 1 7 B1 1 8 B1 0 9 B1 1 10 C1 NA 11 C1 NA 12 C1 NA
Here is a base R solution
dfout <- Reduce(rbind,
lapply(split(data,data$ID),
function(v) {if (!all(is.na(v$Value1))) v$Value1[is.na(v$Value1)]<- 0; v}))
such that
> dfout
ID Value1
1 A1 0
2 A1 2
3 A1 1
4 A1 1
5 B1 0
6 B1 1
7 B1 1
8 B1 0
9 B1 1
10 C1 NA
11 C1 NA
12 C1 NA
With dplyr
:
data %>%
group_by(ID) %>%
mutate(Value1 = ifelse(any(!is.na(Value1)) & is.na(Value1), 0, Value1))
# A tibble: 12 x 2
# Groups: ID [3]
ID Value1
<fct> <dbl>
1 A1 0
2 A1 2
3 A1 1
4 A1 1
5 B1 0
6 B1 1
7 B1 1
8 B1 0
9 B1 1
10 C1 NA
11 C1 NA
12 C1 NA
Using data.table
setDT(data)
data[, Value1 := if (all(is.na(Value1))) NA else replace(Value1, is.na(Value1), 0), by = ID]
ID Value1
1: A1 0
2: A1 2
3: A1 1
4: A1 1
5: B1 0
6: B1 1
7: B1 1
8: B1 0
9: B1 1
10: C1 NA
11: C1 NA
12: C1 NA
The technical post webpages of this site follow the CC BY-SA 4.0 protocol. If you need to reprint, please indicate the site URL or the original address.Any question please contact:yoyou2525@163.com.