Mixed order sorting

[Peter_Griffin]

I am trying to sort a column in a data.frame with mixed orders.
Specifically, I would like to sort it in ascending order first. Then for the highest 30% observations, I want them to be in descending order. How can I do that?

For example,

x <- c(10, 9, 8, 7, 6, 5, 4, 3, 2, 1)

I want to sort it into this:

c(1, 2, 3, 4, 5, 6, 7, 10, 9, 8)

How can I do that?

[technocrat]

To find: a function object f, such that f(x) = y

Solution:

x <- c(10, 9, 8, 7, 6, 5, 4, 3, 2, 1)
y <- c(1, 2, 3, 4, 5, 6, 7, 10, 9, 8)

get_top <- function(x,n) ceiling(quantile(x, probs = n))


make_vct <- function(x,n) {
  c(
    sort(x[x <  ceiling((get_top(x,.7)))]),
    x[x >= ceiling((get_top(x,.7)))])
}

make_vct(x,.7)
#>  [1]  1  2  3  4  5  6  7 10  9  8

^{Created on 2020-09-27 by the reprex package (v0.3.0.9001)}

[Peter_Griffin]

I am trying to make your function work on data.table but couldn't get it done.

mixed_sort <- function(data_table, col, prob)
{data_table1 <- data_table[eval(parse(text = col)) < quantile(eval(parse(text = col)), prob, na.rm = TRUE)]
 setorder(data_table1, eval(parse(text = col)))
 data_table2 <- data_table[eval(parse(text = col)) >= quantile(eval(parse(text = col)), prob, na.rm = TRUE)]
 setorder(data_table1, eval(parse(text = paste0("-", col))))
 return(rbind(data_table1, data_table2))}

data_table <- data.table(x = 1:20)


mixed_sort(data_table, "x", 0.7)

I am not sure how to make this code work.

[technocrat]

What x does the function require, and were is it to be found in data_table?

[Peter_Griffin]

I solved this problem:

mixed_sort <- function(data_table, col, prob)
{data_table1 <- data_table[get(col) < quantile(get(col), prob, na.rm = TRUE)]
data_table1 <- data_table1[order(get(col))]
data_table2 <- data_table[eval(parse(text = col)) >= quantile(eval(parse(text = col)), prob, na.rm = TRUE)]
data_table2 <- data_table2[order(-get(col))]
return(rbind(data_table1, data_table2))}

data_table <- data.table(x = 1:20)

mixed_sort(data_table, "x", 0.7)

[Ajackson]

I do a similar thing as shown below

#   Sort counties with 20 largest first, then alphabetical

ByPop <- arrange(County_pop, -Population)
ByAlpha <- arrange(ByPop[21:nrow(ByPop),], County)
County_pop <- bind_rows(ByPop[1:20,], ByAlpha)
ByPop <- ByAlpha <- NULL