CODEWITHSUNDEEP
rdataframereplacena
Notna
Notna

Reputation: 525

Replace a cell with NA according to value in another cell in R

I have a dataset from which I made a reproducible example:

set.seed(1)
Data <- data.frame(
A = sample(0:5),
B = sample(0:5),
C = sample(0:5),
D = sample(0:5),
corr_A.B = sample(0:5),
corr_A.C = sample(0:5),
corr_A.D = sample(0:5))

> Data
  A B C D corr_A.B corr_A.C corr_A.D
1 1 5 4 2        1        2        4
2 5 3 1 3        5        5        0
3 2 2 3 4        0        1        2
4 3 0 5 0        4        0        1
5 0 4 2 1        2        3        3
6 4 1 0 5        3        4        5

And I would like to check, for each column B, C and D, if one of their cell is equal to 0, I would like to replace, on the same row, the corresponding corr_A column with NA. For instance, since Data$B[4] is equal to 0, I would like Data$corr_A.B[4] to be replaced by NA.

I look to obtain the following result:

> Data
  A B C D corr_A.B corr_A.C corr_A.D
1 1 5 4 2        1        2        4
2 5 3 1 3        5        5        0
3 2 2 3 4        0        1        2
4 3 0 5 0        NA       0       NA
5 0 4 2 1        2        3        3
6 4 1 0 5        3        NA       5

I have tried different ways, using for loops, but I am struggling a lot. Also, in the dataset I am working on, there are many other columns that do not need to be checked for that condition, I would like to be able to specifically designated in which columns I am looking for 0 values.

If someone would be kind enough to give it a try? Many thanks

Upvotes: 0

Views: 1411

Answers (5)

Tim Biegeleisen
Tim Biegeleisen

Reputation: 522762

For a base R solution, we can just use ifelse here:

Data$corr_A.B <- ifelse(Data$B == 0, NA, Data$corr_A.B)
Data$corr_A.C <- ifelse(Data$C == 0, NA, Data$corr_A.C)
Data$corr_A.D <- ifelse(Data$D == 0, NA, Data$corr_A.D)

Upvotes: 2

User7598
User7598

Reputation: 1678

Using apply(). You could do:

cbind(Data,apply(Data[c("B","C","D")],2,function(x){
  ifelse(x==0,NA,x)
}))

Upvotes: 0

Rui Barradas
Rui Barradas

Reputation: 76673

A one-liner using function is.na<-.

is.na(Data[5:7]) <- Data[2:4] == 0

Data
#  A B C D corr_A.B corr_A.C corr_A.D
#1 1 5 4 2        1        2        4
#2 5 3 1 3        5        5        0
#3 2 2 3 4        0        1        2
#4 3 0 5 0       NA        0       NA
#5 0 4 2 1        2        3        3
#6 4 1 0 5        3       NA        5

Upvotes: 5

nicola
nicola

Reputation: 24520

A base, one-liner, vectorized, but convoluted solution:

Data[t(t(which(Data[,2:4]==0,arr.ind=TRUE))+c(0,4))]<-NA

Upvotes: 1

e.matt
e.matt

Reputation: 886

df<- data.frame(A=c(1,5,2,3,0,4),
                B=c(5,3,2,0,4,1),
                C=c(4,1,3,5,2,0),
                D=c(2,3,4,0,1,5),
                corr_A.B=c(1,5,0,4,2,3),
                corr_A.C=c(2,5,1,0,3,4),
                corr_A.D=c(4,0,2,1,3,5))

df %>% mutate(corr_A.B=case_when(B==0 ~ NA_real_,
                                 TRUE~ corr_A.B),
              corr_A.C=case_when(C==0 ~NA_real_,
                                  TRUE ~ corr_A.C),
              corr_A.D=case_when(D==0 ~ NA_real_,
                                  TRUE ~ corr_A.D))
A B C D corr_A.B corr_A.C corr_A.D
1 1 5 4 2        1        2        4
2 5 3 1 3        5        5        0
3 2 2 3 4        0        1        2
4 3 0 5 0       NA        0       NA
5 0 4 2 1        2        3        3
6 4 1 0 5        3       NA        5

Upvotes: 1

Related Questions