Trying to change values in one value depending on values on another variable-CodePudding

I am trying to remove some NA's in one variable depending on the values gotten in a previous variable but I haven't managed to succeed...

Thanks in advance!

There are two variables: info_geo and info_geo_actzd. If the answer in the former one is "No", I want the answers in info_geo_actzd to change from "NA" to "No". The rest of the answers I want to leave them the same as now

I have tried various ways but I can't do it properly

db <- db %>%
    mutate(info_geo_actzd = case_when(
      info_geo == 'No' ~ "No",
      TRUE ~ .))

db <- db %>% 
  mutate(info_geo_actzd=ifelse('info_geo' == 'No' & 'info_geo_actzd' == 'NA', No))

I've also tried this but no result

db <- db  %>%
  mutate(info_geo_actzd = case_when(
    any(info_geo == 'No') ~ 'NA',
    TRUE ~ .))

Thanks in advance

CodePudding user response：

The . will return the dataset instead of the column. If the NAs are unquoted, use is.na to create a logical vector and assign those to "No" or else (TRUE) return the original column info_geo

library(dplyr)
db <- db %>%
    mutate(info_geo_actzd = case_when(
      is.na(info_geo) ~ "No",
      TRUE ~ info_geo))

For more than one column, loop over the columns (either column index or column name - unquoted or quoted, or if there are any patterns in column names, wrap those in select-helpers -contains/matches/starts_with/ends_with etc)

db <- db %>%
    mutate(across(starts_with("info"), ~ case_when(is.na(.x) ~ "No",
         TRUE ~ .x), .names = "{.col}_actzd"))