fshist.r

#!/usr/bin/env r

# create input file with:
# $ find /path/to/dir -type f -printf '%s\n' | gzip > dir.file.sizes.gz
# start with:
# $ zcat dir.file.size.gz | fshist.r 
# edit breaks to your needed sizes
# leave the breaks "0" and "2^50" for the edges

suppressMessages(library(stats))
suppressMessages(library(utils))
suppressMessages(library(gdata))

sizes <- scan("stdin", quiet = T)

h <- hist(sizes, plot = F, breaks = c(0, 1, 3584, 2^15, 2^18, 2^20, 2^24, 2^30, 2^37, 2^50))

d <- data.frame (
  min = humanReadable(head(h$breaks, -1)),
  max = humanReadable(tail(h$breaks, -1)),
  count = h$counts
)

print(d)