CODEWITHSUNDEEP
rdplyr
rais
rais

Reputation: 103

Selecting large number of columns for recoding the variable values

I am trying to recode a group of variables in a dataframe(wide format). Let's say in that dataframe, I have a variable - "Education status" which runs from month 1 to 150, labeled as EDU1, EDU2, EDU3 .... EDU150 across 150 columns. The variable has values = 0,1,2,3,4,5,NA. I am trying to recode the variables (0=91,1=92,2=93,3,4,5=94) while keeping NA as it is. Since it is a large number of columns, I am trying to do it at once. This is what I tried.

First selecting columns from EDU1 to EDU150 at once.

library(dplyr)
df %>%
 mutate_at(c(EDU1:EDU150),
 ~recode(.,"0=91;1=92,2=93;3=94;4=94;5=94"))

Another way I tried is picking up a part of the column name i.e starts with 'EDU'and specifying where to run the recode function.

library(dplyr)
df%>% 
mutate(across(starts_with('EDU'), ~ recode(.,"0=91;1=92,2=93;3=94;4=94;5=94")))

None of this work, and I get an error saying

Unreplaced values treated as NA as.x is not compatible. Please specify replacements exhaustively or supply .default

Can anyone help me with this? Is it because I have some NA values in the variables I am working at or something is wrong with the code. What would be the best way to run a function selecting a large number of columns by picking up the column name?

Upvotes: 1

Views: 455

Answers (1)

forhad
forhad

Reputation: 196

library(dplyr)
df %>%
  mutate(across(
    EDU1:EDU150,
    ~ recode(
      .x,
      `0` = 91L,
      `1` = 92L,
      `2` = 93L,
      `3` = 94L,
      `4` = 94L,
      `5` = 94L
    )
  ))

Upvotes: 1

Related Questions