Re: --min-size option for du (patch)

[Brock Noland]

Hi,

Not sure if this patch will be accepted or not, just wondering why awk
won't work for this purpose? A quick example is below. I would think
awk combined with sort could provide some very nicely formatted
output.

# du -sb * | awk '$1 > 1024*1024 {print $1/(1024*1024) "MB", $2}'
7.5757MB 200707
26.1926MB addons.tar.gz
3.03811MB awstats-6.7
111.746MB backup.tgz
11.1758MB bootcpy

Thanks,
Brock

On Wed, Jan 14, 2009 at 4:54 PM, Paul Chambers <address@hidden> wrote:
> I'd like to submit the attached small patch for du. it adds a '--min-size'
> option, which can be used to filter the output to only include items of that
> size or greater.
>
> I find it useful in locating large folders on disk in one pass, to zero in
> on the worst offenders when disk space is getting low.
>
> regards,
>
> -- Paul
>
> diff -ru coreutils-6.12/man/du.1 coreutils-6.12p/man/du.1
> --- coreutils-6.12/man/du.1     2008-05-31 14:26:01.000000000 -0700
> +++ coreutils-6.12p/man/du.1    2008-11-09 09:50:24.000000000 -0800
> @@ -1,5 +1,5 @@
>  .\" DO NOT MODIFY THIS FILE!  It was generated by help2man 1.35.
> -.TH DU "1" "May 2008" "GNU coreutils 6.12" "User Commands"
> +.TH DU "1" "November 2008" "GNU coreutils 6.12" "User Commands"
>  .SH NAME
>  du \- estimate file space usage
>  .SH SYNOPSIS
> @@ -90,6 +90,10 @@
>  line argument;  \fB\-\-max\-depth\fR=\fI0\fR is the same as
>  \fB\-\-summarize\fR
>  .TP
> +\fB\-\-min\-size\fR=\fISIZE\fR
> +Only print the total for a directory (or file, with
> +\fB\-\-all\fR) if it is equal to or greater than this value.
> +.TP
>  \fB\-\-time\fR
>  show time of the last modification of any file in the
>  directory, or any of its subdirectories
> diff -ru coreutils-6.12/src/du.c coreutils-6.12p/src/du.c
> --- coreutils-6.12/src/du.c     2008-05-27 23:29:58.000000000 -0700
> +++ coreutils-6.12p/src/du.c    2008-11-09 09:50:18.000000000 -0800
> @@ -153,6 +153,10 @@
>    is at level 0, so `du --max-depth=0' is equivalent to `du -s'.  */
>  static size_t max_depth = SIZE_MAX;
>
> +/* Only show the total for each directory (and file if --all) if it is
> +       equal to or greater than this value.  */
> +static uintmax_t min_size = 0;
> +
>  /* Human-readable options for output.  */
>  static int human_output_opts;
>
> @@ -198,6 +202,7 @@
>   FILES0_FROM_OPTION,
>   HUMAN_SI_OPTION,
>   MAX_DEPTH_OPTION,
> +  MIN_SIZE_OPTION,
>   MEGABYTES_LONG_OPTION,
>   TIME_OPTION,
>   TIME_STYLE_OPTION
> @@ -218,6 +223,7 @@
>   {"human-readable", no_argument, NULL, 'h'},
>   {"si", no_argument, NULL, HUMAN_SI_OPTION},
>   {"max-depth", required_argument, NULL, MAX_DEPTH_OPTION},
> +  {"min-size", required_argument, NULL, MIN_SIZE_OPTION},
>   {"null", no_argument, NULL, '0'},
>   {"no-dereference", no_argument, NULL, 'P'},
>   {"one-file-system", no_argument, NULL, 'x'},
> @@ -322,6 +328,8 @@
>                           only if it is N or fewer levels below the
> command\n\
>                           line argument;  --max-depth=0 is the same as\n\
>                           --summarize\n\
> +      --min-size=SIZE   Only print the total for a directory (or file,
> with\n\
> +                          --all) if it is equal to or greater than this
> value.\n\
>  "), stdout);
>       fputs (_("\
>       --time            show time of the last modification of any file in
> the\n\
> @@ -608,8 +616,9 @@
>   if (!print)
>     return ok;
>
> -  if ((IS_DIR_TYPE (ent->fts_info) && level <= max_depth)
> -      || ((opt_all && level <= max_depth) || level == 0))
> +  if ((IS_DIR_TYPE (ent->fts_info) || opt_all)
> +      && (level <= max_depth || level == 0)
> +         && (dui_to_print.size >= min_size))
>     print_size (&dui_to_print, file);
>
>   return ok;
> @@ -773,6 +782,15 @@
>          }
>          break;
>
> +       case MIN_SIZE_OPTION:           /* --min-size=SIZE */
> +         {
> +               enum strtol_error e = xstrtoumax (optarg, NULL, 0,
> &min_size,
> +                                      "eEgGkKmMpPtTyYzZ0");
> +               if (e != LONGINT_OK)
> +             xstrtol_fatal (e, oi, c, long_options, optarg);
> +      }
> +         break;
> +
>        case MEGABYTES_LONG_OPTION:
>          error (0, 0,
>                 _("the --megabytes option is deprecated; use -m instead"));
>
> _______________________________________________
> Bug-coreutils mailing list
> address@hidden
> http://lists.gnu.org/mailman/listinfo/bug-coreutils
>
>