如何将水平行调整为跨大数据集的多行

Question

We have a large dataset which we would like to edit and analyse but before we can begin we need to transpose the data into a more functional format for statistical analysis.

````Incorrect format dataframe
library(tidyverse)
data <-
 tribble(~id, ~date, ~start, ~end, ~start, ~end, ~start, ~end,
         1001, "01/07/2019", "04:00", "08:00", "10:00", "15:00", "16:00", "20:00",
         1001, "02/07/2019", "04:30", "05:30", "09:00", "14:00", "17:00", "21:00",
         1009, "05/07/2019", "03:00", "05:00", "07:00", "14:00", "15:00", "19:00",
         1009, "07/07/2019", "03:30", "04:30", "08:20", "15:20", "16:30", "20:30") 

````Correct format dataframe
# id date start end
# 1001 01/07/2019 04:00 08:00
# 1001 01/07/2019 10:00 15:00
# 1001 01/07/2019 16:00 20:00
# 1001 02/07/2019 04:30 05:30
# 1001 02/07/2019 09:00 14:00
# 1001 02/07/2019 17:00 21:00
# 1009 05/07/2019 03:00 05:00
# 1009 05/07/2019 07:00 14:00
# 1009 05/07/2019 15:00 19:00
# 1009 07/07/2019 03:30 04:30
# 1009 07/07/2019 08:20 15:20
# 1009 07/07/2019 16:30 20:30

I can manipulate my data manually but I've been unable to conduct automated function. The actual dataset has 32 columns across and 10,000 rows. Edit: I've tried to concatenate id and date to every value and sort, but have made mistakes with this method.

Answer 1

Next time it would be great if you could poste a reproducible example of your data (like the one in my code below).

It looks like what you want to do is to turn your data from a wide into some kind of long format. The duplicated column names are causing some trouble but the code below should do the trick. You will have to install the tidyverse package for this:

library(tidyverse)
  data <-
     tribble(~id, ~date, ~start, ~end, ~start, ~end, ~start, ~end,
             1001, "01/07/2019", "04:00", "08:00", "10:00", "15:00", "16:00", "20:00",
             1001, "02/07/2019", "04:30", "05:30", "09:00", "14:00", "17:00", "21:00",
             1009, "05/07/2019", "03:00", "05:00", "07:00", "14:00", "15:00", "19:00",
             1009, "07/07/2019", "03:30", "04:30", "08:20", "15:20", "16:30", "20:30") 
  # make column names unique
  names(data) <-
     ifelse(names(data) %in% c("start","end"),
            paste0(names(data),"_",1:length(names(data))),
           names(data))

  # turn data into long format
  data %>%
     gather(key,value,-id,-date) %>%
     arrange(id,date) %>%
     # get rid of the column suffixes
     mutate(key = str_replace_all(key,pattern = c("_\\d+"=""))) %>% 
     group_by(id,date,key) %>% 
     mutate(obs_id = row_number()) %>% 
     spread(key,value) %>% 
     ungroup() %>% 
     select(id,
            date,
            start,
            end)