Utilities for finding large or unwanted files/dirs.

find-bigdirs

#! /bin/bash

# Find big dirs, minimally deep.
# Running in near-silent mode in case want to pipe output.
#
# Author: Micah Elliott <mde MicahElliott com>

#dt=$(date "+%Y%m%d")
#log="bigdirs.$dt.log"
log="$PWD/bigdirs.log"
dir=${1:-$PWD}

echo "Finding big dirs in “$dir”; results in “$log”" 1>&2

# Will assume ‘.’ in search for cleaner output paths.
cd "$dir"

# See semblance of result for debugging.
#du  *  --max-depth 1 --one-file-system | sort -nr

### Find dirs bigger than ~40 MiB.
# Throwing away some nice size output, but this makes the processing
# simple.
du  *  --max-depth 1 --one-file-system \
| sort -nr \
| egrep -e '^[4-9][[:digit:]]{4}.*' -e '^[[:digit:]]{6,}.*' \
| grep '/' \
| cut -f2 \
| tee $log

find-bigfiles

#! /bin/bash

# Find unwanted big files (and misplaced archives).
#
# Author: Micah Elliott <mde MicahElliott com>

#dt=$(date "+%Y%m%d")
#log="bigfiles.$dt.log"
log="$PWD/bigfiles.log"
dir=${1:-$PWD}

echo "Finding big files in “$dir”; results in “$log”" 1>&2

# Will use ‘.’ in find for cleaner paths.
cd "$dir"

# Find files bigger than 10 MiB (10,000,000 bytes)
find . \
-type f -a  \( \
   -size +10000000c -a \( \
      -name  '*.so' -o \
      -name  '*.a' -o \
      -name  '*.x' -o \
      -name  '*.exe' -o \
      -name  '*.o' -o \
      -name  '*.dat' -o \
      -name  '*.in' \
   \) \
   -o \( \
      -iname '*.rpm' -o \
      -iname '*.tar' -o \
      -iname '*.gz' -o \
      -iname '*.tgz' -o \
      -iname '*.taz' -o \
      -iname '*.bz2' -o \
      -iname '*.Z' -o \
      -iname '*.iso' -o \
      -name  'core' -o \
      -iname '*.bin' -o \
      -name  '*~' -o \
      -name  '*.sw[po]' -o \
      -iname '*.zip' -o \
      -iname '*.gzip' \
      -iname '*.bak' \
      -iname '*.orig' \
      -name  '*.pyc' \
   \) \
\) \
-printf "%P\n" \
| tee $log

find-tmps

#! /bin/zsh

# Find temporary files (and unwanted archives).
#
# Author: Micah Elliott <mde MicahElliott com>

dir=${1:-$PWD}

tmptypes=(
    '*.so'
    '*.a'
    '*.x'
    '*.exe'
    '*.o'
    '*.dat'
    '*.in'
    '*.rpm'
    '*.tar'
    '*.gz'
    '*.tgz'
    '*.taz'
    '*.bz2'
    '*.Z'
    '*.iso'
    'core'
    '*.bin'
    '*~'
    '*.sw[po]'
    '*.zip'
    '*.gzip'
    '*.bak'
    '*.orig'
    '*.pyc'
)

#patterns=()
for t in $tmptypes; do
    patterns+=" -name $t -o "
done
patterns+=" -name DOES_NOT_EXIST "

echo "Finding temporary files." 1>&2

cd "$dir"

# Surely there is a better way to expand $patterns, but this is the
# only way I’m finding that actually works. Other attempts
# accidentally expand the globs above, or expand $patterns as a quoted
# param.
find . -type f -a \( $(echo "$patterns") \) -printf "%P\n"

# Sort numerically .  This needs to be piped into from above 'find'.
#ls -l $(cat $file) | sed -e 's/[ ]\+/ /g' | cut -f5,9 -d" " | sort -nr > $file