Output from the function in the desired format-CodePudding

my data

    data <- structure(list(col1 = 1:9, col2 = 10:18, col3 = 16:24, col4 = 67:75, 
    col5 = c(19L, 19L, 19L, 19L, 19L, 19L, 19L, 19L, 19L), GROUP = c(1L, 
    1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L)), class = "data.frame", row.names = c(NA, 
-9L))

Func:

    combination <- list(c(1, 2), c(1, 3), c(1, 4),c(2,3),c(2,4),c(2,5),c(3,4),c(3,5))

wilcox.fun <- function(dat) { 
  do.call(rbind, lapply(combination, function(x) {
    test <- wilcox.test(dat[[x[1]]], dat[[x[2]]], paired=TRUE)
    data.frame(Test = sprintf('%s by %s', x[1],x[2]), 
               #W = round(test$statistic,4),
               med = paste(median(dat[[x[1]]]),median(dat[[x[2]]])),
               p = test$p.value)
  }))
}

result <- purrr::map_df(split(data, data$GROUP), wilcox.fun, .id = 'Group')

I have a function that takes values from pairwise combinations and finds med and p. I am not satisfied with the format in which it outputs the result.what I want to get:

resulte <- structure(list(Group = c(1L, 1L, 1L, 1L, 1L), Test = 1:5, med = c(5L, 
14L, 20L, 71L, 19L), p = c("1 by 2: 0,00335343645494632\n1 by 3: 0,00335343645494632\n1 by 4: 0,00335343645494632", 
"2 by 3: 0,00335343645494632\n2 by 4: 0,00335343645494632\n2 by 5: 0,00390625\n", 
"3 by 4: 0,00335343645494632\n3 by 5: 0,325204163250902", NA, 
NA)), class = "data.frame", row.names = c(NA, -5L))

CodePudding user response：

Maybe you are looking for map_dfr? I also edited the function a bit:

library(tidyverse)

data <- structure(list(col1 = 1:9, col2 = 10:18, col3 = 16:24, col4 = 67:75, 
    col5 = c(19L, 19L, 19L, 19L, 19L, 19L, 19L, 19L, 19L), GROUP = c(1L, 
    1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L)), class = "data.frame", row.names = c(NA, 
-9L))

#add another group here for example
data2 <- bind_rows(data, data |> mutate(GROUP = 2))

wilcox.fun <- function(dat){
  t(combn(paste0("col", 1:5), 2)) |>
    as.data.frame() |>
    mutate(test = map2(V1, V2, 
                       \(x, y) wilcox.test(dat[,x], dat[,y], paired=TRUE)),
           p = map_dbl(test, \(x) x$p.value ),
           Test = glue::glue("{V1} by {V2}")) |>
    select(Test, p)
}


map_dfr(split(data2, data2$GROUP), wilcox.fun, .id = 'Group')
#>    Group         Test           p
#> 1      1 col1 by col2 0.003353436
#> 2      1 col1 by col3 0.003353436
#> 3      1 col1 by col4 0.003353436
#> 4      1 col1 by col5 0.003906250
#> 5      1 col2 by col3 0.003353436
#> 6      1 col2 by col4 0.003353436
#> 7      1 col2 by col5 0.003906250
#> 8      1 col3 by col4 0.003353436
#> 9      1 col3 by col5 0.325204163
#> 10     1 col4 by col5 0.003906250
#> 11     2 col1 by col2 0.003353436
#> 12     2 col1 by col3 0.003353436
#> 13     2 col1 by col4 0.003353436
#> 14     2 col1 by col5 0.003906250
#> 15     2 col2 by col3 0.003353436
#> 16     2 col2 by col4 0.003353436
#> 17     2 col2 by col5 0.003906250
#> 18     2 col3 by col4 0.003353436
#> 19     2 col3 by col5 0.325204163
#> 20     2 col4 by col5 0.003906250

EDIT

Here is some code to get the requested format:

library(tidyverse)

result <- tibble(Group = unique(data$GROUP),
       Test = 1:length(colnames(data)[grepl("col", colnames(data))]),
       med = map2_dbl(Test, Group, \(x, y) median(data[data$GROUP == y,x])),
       p = map2_chr(Test, Group, \(x,y){
         my_dat <- data[data$GROUP == y,]
         as.data.frame(t(combn(paste0("col", 1:5), 2))) |>
            filter(V1 == paste0("col", x))|>
              mutate(test = map2(V1, V2, 
                                 \(q, r) wilcox.test(my_dat[,q], my_dat[,r], paired=TRUE)),
                     p = map_dbl(test, \(j) j$p.value ),
                     out = paste0(x, " by ", {min(c(x 1,5))}:5, ": ", p, "\n")) |>
              pull(out) |>
              paste(collapse = "")
       }))