How to replace non-positive values of some columns of my dataframe with NA in r-CodePudding

I have a dataframe, and I want to assign NA to non-positive values of specific columns.

I'll try to realize a minimal reproducible example with the mtcars dataframe where I'll try to change with NA the values of the two columns cyl and disp that are <= 5.

library(dplyr)

view(mtcars)


Nomi <- c("cyl", "disp")

for(i in Nomi) {
  mtcars$i[mtcars$i <= 5] <- NA
}

Nomi is the vector with the names of mtcars columns where I want to change values with NA.

I don't fully understand the error message I get, I'd like to find a solution that can make the changes in NA.

CodePudding user response：

In base R, you can directly do the replacement on the dataframe.

Nomi <- c("cyl", "disp")
df <- mtcars
df[Nomi][df[Nomi] <= 5] <- NA
df

CodePudding user response：

If you are comfortable with dplyr, you can use an ifelse statement in mutate(across()). It means across cyl and disp column, if the values in these two columns are <= 5, replace it with NA. Otherwise, keep the original value (.x).

Note that this will only output results to the console WITHOUT actually replacing the mtcars dataset.

library(dplyr)

mtcars %>% mutate(across(c(cyl, disp), ~ ifelse(.x <= 5, NA, .x)))