Change all numerical values to NA in row based on first column values in R-CodePudding

I have a huge df and a simplified version of my problem is:

A <- c("CA1", "CA2", "CA3", "CA4")
B <- c( 0,1,0,1)
C <- c(1,1,0,0)

df <- data.frame(A, B, C)

I would like all row values that correspond to CA3 and CA4 to change to NA and the resulting df2:

A     B   C
CA1   0   1
CA2   1   1
CA3   NA  NA
CA4   NA  NA

Is it possible to do this with base R and dplyr? The code I have adapted but of course isn't working is:

df2 <- df %>% mutate_at(row("CA3", "CA4"), na_if, 0) %>%
mutate_at(row("CA3", "CA4"), na_if, 1)

my error code is: Error in if (as.factor) { : argument is not interpretable as logical

Any ideas on how to generate a code to get the df2 output?

Thank you!

CodePudding user response：

Loop across the 'B', 'C' column, create the condition where the 'A' value is 'CA3', 'CA4' with %in% and change the values in case_when. The default TRUE condition in case_when returns NA if we specify the other condition to return the original column values

library(dplyr)
df2 <- df %>% 
    mutate(across(B:C, ~ case_when(!A %in% c("CA3", "CA4") ~ .)))

-output

df2
    A  B  C
1 CA1  0  1
2 CA2  1  1
3 CA3 NA NA
4 CA4 NA NA

CodePudding user response：

Accomplishing the same task in base R:

df[df$A %in% c("CA3", "CA4"), c("B", "C")] <- NA
df
    A  B  C
1 CA1  0  1
2 CA2  1  1
3 CA3 NA NA
4 CA4 NA NA

CodePudding user response：

Here is a pivoting alternative:

library(dplyr)
librayr(tidyr)

df %>% 
  pivot_longer(
    cols = -A
  ) %>% 
  mutate(value = ifelse(A == "CA3" | A == "CA4", NA, value)) %>% 
  pivot_wider(
    names_from = name
  )

  A         B     C
  <chr> <dbl> <dbl>
1 CA1       0     1
2 CA2       1     1
3 CA3      NA    NA
4 CA4      NA    NA