finding top 1 percentile in datatable in r

Question

I have a data table like

     sample1 sample2 sample3

fruit1    10      20      30

fruit2     1      5       6

fruit3     3      7       8

etc.

I want to find the top 1 percentile of fruits in each sample in R (according to the number in each sample). Is there a simple way to do this?

Answer 1

You can lapply over your data and for each column, subset the rownames of df with a logical vector which is TRUE when the corresponding value in the column is in the 1 percentile (ie above the 100 - 1 percentile).

Create example data

set.seed(2019)
df <- as.data.frame(matrix(sample(1e4, replace = T), 1e3, 10))
names(df) <- paste0('sample', seq_along(df))
rownames(df) <- paste0('fruit', seq_len(nrow(df)))

Step described above:

lapply(df, function(x) rownames(df)[x > quantile(x, (100 - 1)/100)])

# $`sample1`
#  [1] "fruit57"  "fruit76"  "fruit149" "fruit471" "fruit520" "fruit682" "fruit805"
#  [8] "fruit949" "fruit966" "fruit975"
# 
# $sample2
#  [1] "fruit49"  "fruit109" "fruit232" "fruit274" "fruit312" "fruit795" "fruit883"
#  [8] "fruit884" "fruit955" "fruit958"
# 
# $sample3
#  [1] "fruit37"  "fruit189" "fruit231" "fruit256" "fruit473" "fruit654" "fruit729"
#  [8] "fruit742" "fruit820" "fruit979"
# 
# ...

Answer 2

Assuming your data frame is calle "fruit"

fruit <- fruit[order(fruit$sample1,decreasing = TRUE)]

top.1.percent <- fruit[1:length(fruit$sample1)/100,]

This should do the trick for sample1