CODEWITHSUNDEEP
r
Aurélien
Aurélien

Reputation: 113

R aggregate multiple column in a unique

I have in my data frame columns for differents regions like this :

data.frame(value=c(1,2,3,4,5),europe=c("yes","no","no","no","no"),
am_north=c("no","yes","no","no","no"),am_south=c("no","no","yes","no","no"),moaf=c("no","no","no","yes","no"),asia=c("no","no","no","no","yes"))

I would like to remove all region column and keep a new column named "region" with the name of the region like that :

data.frame(value=c(1,2,3,4,5),region=c("europe","am_north","am_south","moaf","asia"))

I know there is a package for that like tidy or reshape, but i don't find to do it like this in a easy and simple way

Upvotes: 0

Views: 69

Answers (3)

denis
denis

Reputation: 5673

and with data.table

melt(DT, id.vars = "value",value.name = "answer",variable.name = "region")[answer == "yes",.(value,region)]
   value   region
1:     1   europe
2:     2 am_north
3:     3 am_south
4:     4     moaf
5:     5     asia

Upvotes: 0

Deena
Deena

Reputation: 6223

Using base-r: 

df[["region"]] = colnames(df)[apply(df == "yes", 1, which)]
df[, c("value", "region")]
# value   region
# 1     1   europe
# 2     2 am_north
# 3     3 am_south
# 4     4     moaf
# 5     5     asia

Upvotes: 2

Jake Kaupp
Jake Kaupp

Reputation: 8072

If you're wanting to do this to arrange the data based off of different conditions in the data frame, then here is a tidyverse approach.

library(tidyverse)

gather(df, region, response, -value) %>% 
  filter(response == "yes") %>% 
  select(-response)

  value   region
1     1   europe
2     2 am_north
3     3 am_south
4     4     moaf
5     5     asia

Upvotes: 3

Related Questions