Bash script for creating a list with largest files

A few comments about your getopts usage:

1. You are duplicating some of what getopts does by checking $@ before you even use getopts. It's more common (and, IMO, better) to set some defaults for your option variables, then use getopts to allow the user to override the defaults.

2. Do you really want all those warnings about using default values? It's good to warn about actual errors (like not providing a mandatory option where no sensible default is possible), but most users would consider a warning about perfectly normal things like a program using default values to be annoying noise. The defaults can be printed in the HelpFunction message.

3. Your script doesn't use any non-option arguments (like filenames), but in scripts that do, you will need to use shift to remove the options (and shift again to remove their OPTARG(s) if any) from the argument list. It wouldn't be any use to do that in this script, but it wouldn't hurt to do it either (and it's good habit / good practice).

And general comments:

1. Double-quote your variable expansions. You mostly do this, which is good. Sometimes you don't, which is not. See Why does my shell script choke on whitespace or other special characters?

2. Use printf and/or heredocs instead of echo, especially when printing variables. See Why is printf better than echo?

3. Don't parse the output of ls. ls's output is for viewing by human eyes, not for consumption by programs. See Why not parse ls (and what to do instead)?.

   There are two main circumstances where you're using the output of ls in your script.

   • Testing if a file exists. Bash can already do this easily for single files with test/[ ... ] or [[ ... ]].

     For wildcards, it's a bit more complicated but you still shouldn't use ls. Use find instead - that's what it's for. The trouble is that find always returns an exit code of 0 unless an error occurs. We can, however, use find and grep to check if at least one match was found.

     For example, don't do this:

     if ls $DATA_DIR/$HLQ_DIR/*$SEARCH_STRING* 1> /dev/null 2>$1; then
     

     (BTW, there are two bugs on that line: unquoted variables and 2>$1 - that's probably a typo for 2>&1)

     Do this instead:

     if find "$DATA_DIR/$HLQ_DIR/" -maxdepth 1 -type f \
          -name "*$SEARCH_STRING*" 2>/dev/null |
        grep -q '.' ; then
     

     If there aren't any errors (e.g. permissions), this shouldn't produce any output, but I'm redirecting stderr to /dev/null in case there are.

     BTW, find has -regex and -iregex (case-insensitive) options to search with regular expressions rather than glob patterns. Also a -regextype option to set what "dialect" of regular expressions to use.

   • Extracting one or more filenames into a scalar variable or an array. Using ls for this is never a good idea. Either use a glob or use find. E.g. Instead of using ls and command substitution like this:

     OUTPUT=$(ls -1fsh $DATA_DIR/$HLQ_DIR/*$SEARCH_STRING* 2> /dev/null | sort -rh | head -$NUMBER_RESULT)
     

     Use mapfile (AKA readarray) and process substitution to populate an array called OUTPUT by running find.

     The following uses GNU find's -printf to print each file's size (in an 8-character wide field), another 3 spaces, its pathname, and a NUL. Output from find is then piped into the GNU versions sort, head, and numfmt, all of which can handle NUL-separated input with the -z option:

     mapfile -d '' -t OUTPUT < <(
       find "$DATA_DIR/$HLQ_DIR/" \
         -maxdepth 1 \
         -name "*SEARCH_STRING*" \
         -printf '%8s   %p\0'|
       sort -z -r -n -k1,1 |
       head -z -n "$NUMBER_RESULT" | numfmt -z --to=si)
     

     However, see below - this entire block of code (under the #Get basedirecory including subdir "/data" comment) can be improved:

(extra newlines added for readability and to avoid horizontal scroll bars on this site)

# Get base directory including subdir "/data"
mapfile -d '' -t ARRAY_DATA_DIR < <(
  find "$SHARED_BASE_DIR/$TENANT" \
    -maxdepth 3 -type d -name 'data' -print0)

for DATA_DIR in "${ARRAY_DATA_DIR[@]}"; do
    # Get all sub directories using GNU find's -printf %f (basename)
    mapfile -d '' -t ARRAY_HLQ_DIR < <(
      find "$DATA_DIR" -mindepth 1  -maxdepth 1 \
        -type d -printf '%f\0')

    for HLQ_DIR in "${ARRAY_HLQ_DIR[@]}"; do
      # get $NUMBER_RESULT largest files in $HLQ_DIR
      mapfile -d '' -t OUTPUT < <(
        find "$DATA_DIR/$HLQ_DIR/" \
          -maxdepth 1 -type f  -name "*$SEARCH_STRING*" \
          -printf '%8s   %p\0' |
        sort -z -r -n -k1,1 |
        head -z -n "$NUMBER_RESULT" | numfmt -z --to=si)

      # print formatted output with bash's built-in printf
      if [ -n "$OUTPUT" ]; then  # we have at least one file
         printf  "%8s   %s\n" '[ Size ]' "[ --- $HLQ_DIR --- ]";
         printf "%s\n" "${OUTPUT[@]}" "$LNE"
      fi
    done
done

This example:

• quotes all variable expansions;
• never parses ls;
• never uses whitespace as a filename separator, always uses NUL. It is reliable and safe to use with any valid filename;
• uses mapfile to populate arrays with find's output. -d '' ensures it also uses NUL as the separator;
• uses Bash's built-in printf instead of piping to awk.

PS: I just noticed that this question is over 6 months old. Oh well, I've written my answer now and hopefully @JAK will come back to see it. If not, maybe someone else will find something useful in it.