I am tidying my data in R, and want to turn multiple columns into 1, using a function iterating over the items of a vector. I was wondering whether you could help me out to:
My data is based on a survey with 32 questions. Each question has multiple answers. Each answer is a column, with options 1 and NA.
For one question, a section of the dataset can be reproduced as follows:
XV2_1 <- c(1,NA,NA,NA)
XV2_2 <- c(NA,1,NA,NA)
XV2_3 <- c(NA,NA,NA,1)
XV2_4 <- c(NA,NA,1,NA)
id <- c(12,13,14,15)
dat <- data.frame(id,XV2_1, XV2_2, XV2_3,XV2_4)
> dat
id XV2_1 XV2_2 XV2_3 XV2_4
1 12 1 NA NA NA
2 13 NA 1 NA NA
3 14 NA NA NA 1
4 15 NA NA 1 NA
This is the data I would like to have (
question_2_answers <- c("Yellow","Blue","Green","Orange") #this is a vector based on the answers of the questionnaire
collapsed <- c("Yellow","Blue","Orange","Green")
collapsed_dataframe <- data.frame(id,collapsed)
>collapsed_dataframe
id X2
1 12 Yellow
2 13 Blue
3 14 Green
4 15 Orange
So far, I tried a sequence of "ifelse's" combined with mutate:
library(tidyverse)
question_2_answers <- c("Yellow","Blue","Green","Orange") #this is a vector based on the answers of the questionnaire
dat %>%
mutate(
Colour = tidy_Q2(question_2_answers,XV2_1,XV2_2,XV2_3,XV2_4)
)
tidy_Q2 <- function(a,b,c,d,e) {
ifelse(b == 1, a[1],ifelse(
c==1,a[2],ifelse(
d==1,a[3],a[4])))
}
However, my output is not as expected:
id XV2_1 XV2_2 XV2_3 XV2_4 Colour
1 12 1 NA NA NA Yellow
2 13 NA 1 NA NA <NA>
3 14 NA NA NA 1 <NA>
4 15 NA NA 1 NA <NA>
I would have liked it to be as follows:
id XV2_1 XV2_2 XV2_3 XV2_4 Colour
1 12 1 NA NA NA Yellow
2 13 NA 1 NA NA Blue
3 14 NA NA NA 1 Green
4 15 NA NA 1 NA Orange
Does anyone know a way to remove the error? Another question that I'd like to ask, is whether my code can be more efficient? I have 32 survey_questions in store after this, I'd like to automate the process as much as possible. Notable things to take in mind:
Always happy to learn,
Best,
Maria
This is a kind of wide-to-long conversion which we can do with tidyr::gather
:
First, we make the colors the column names of the appropriate rows:
# Replace column names (except for the `id` column) with color values
colnames(dat)[-1] <- c("Yellow","Blue","Orange","Green")
dat
id Yellow Blue Orange Green
1 12 1 NA NA NA
2 13 NA 1 NA NA
3 14 NA NA NA 1
4 15 NA NA 1 NA
Then, we gather the non-id columns and drop the NA values:
library(tidyverse)
dat %>%
gather(X2, val, -id) %>% # Gather color cols from wide to long format
filter(!is.na(val)) %>% # Drop rows with NA values
select(-val) # Remove the unnecessary `val` column
id X2
1 12 Yellow
2 13 Blue
3 15 Orange
4 14 Green
This will work with any number of columns (you just need to specify all columns you don't want to gather) and keeps rows with non- NA
values. If you want other conditions to exclude a row (for example, if 0
or 'unknown'
should count as a non-answer, or only 'correct'
counts as an answer) then you should add those conditions to the filter
statement.
One option in base R
would be max.col
is to find the column index of values that are not NA
in each row, use that to get the column names corresponding to the index, create a 2 column data.frame by cbind
ing with the first column
i1 <- max.col(!is.na(dat[-1]), 'first')
cbind(dat['id'], Colour = names(dat)[-1][i1])
# id Colour
#1 12 Yellow
#2 13 Blue
#3 14 Green
#4 15 Orange
dat <- structure(list(id = c(12, 13, 14, 15), Yellow = c(1, NA, NA,
NA), Blue = c(NA, 1, NA, NA), Orange = c(NA, NA, NA, 1), Green = c(NA,
NA, 1, NA)), class = "data.frame", row.names = c(NA, -4L))
The technical post webpages of this site follow the CC BY-SA 4.0 protocol. If you need to reprint, please indicate the site URL or the original address.Any question please contact:yoyou2525@163.com.