[BASH] Convert bytes to human readable

bytesToHR.sh

function bytesToHR()
{
  local SIZE=$1
  local UNITS="B KiB MiB GiB TiB PiB"
  for F in $UNITS; do
    local UNIT=$F
    test ${SIZE%.*} -lt 1024 && break;
    SIZE=$(echo "$SIZE / 1024" | bc -l)
  done

  if [ "$UNIT" == "B" ]; then
    printf "%4.0f    %s\n" $SIZE $UNIT
  else
    printf "%7.02f %s\n" $SIZE $UNIT
  fi
}

bytesToHR 1
bytesToHR 1023
bytesToHR 1024
bytesToHR 12345
bytesToHR 123456
bytesToHR 1234567
bytesToHR 12345678

This script outputs:

   1    B
1023    B
   1.00 KiB
  12.06 KiB
 120.56 KiB
   1.18 MiB
  11.77 MiB

#!/usr/bin/env bash

function hr() {
    local S="$1"
    for U in B KiB MiB GiB TiB PiB EiB ZiB; do
        test "$S" -lt 1024 && echo "$S$U" && break;
        S="$((S/1024))"
    done
}

Thanks however your version will truncate the result of each division, resulting in a less accurate result.

Maybe, but it doesn't seem to be an issue in practice:

$ hr $((1024*1024*1024*1024))
1TiB

$ hr $((1024*1024*1024*1024-1))
1023GiB

This seems right, though its reasonable if you'd prefer it to round up in this case.

Do you have a particular case in mind where it would be a mistake?

It doesn't replicate the original output with two decimal places of accuracy or aligned formatting.

That's true.

In my use case (printing out the approximate size of a directory in ci) I prefer fewer lines to formatted or highly precise output.

Thanks for sharing your original version btw, this was the best design I found that didn't use numfmt which is missing from github windows runners.