15.2. Complex Commands

Commands for more advanced users

find

-exec COMMAND \;

Carries out COMMAND on each file that find matches. The command sequence terminates with ; (the ";" is escaped to make certain the shell passes it to find literally, without interpreting it as a special character).

 bash$ find ~/ -name '*.txt'
 /home/bozo/.kde/share/apps/karm/karmdata.txt
 /home/bozo/misc/irmeyc.txt
 /home/bozo/test-scripts/1.txt
 	      

If COMMAND contains {}, then find substitutes the full path name of the selected file for "{}".

   1 find ~/ -name 'core*' -exec rm {} \;
   2 # Removes all core dump files from user's home directory.

   1 find /home/bozo/projects -mtime 1
   2 #  Lists all files in /home/bozo/projects directory tree
   3 #+ that were modified within the last day.
   4 #
   5 #  mtime = last modification time of the target file
   6 #  ctime = last status change time (via 'chmod' or otherwise)
   7 #  atime = last access time
   8 
   9 DIR=/home/bozo/junk_files
  10 find "$DIR" -type f -atime +5 -exec rm {} \;
  11 #                                      ^^
  12 #  Curly brackets are placeholder for the path name output by "find."
  13 #
  14 #  Deletes all files in "/home/bozo/junk_files"
  15 #+ that have not been accessed in at least 5 days.
  16 #
  17 #  "-type filetype", where
  18 #  f = regular file
  19 #  d = directory
  20 #  l = symbolic link, etc.
  21 #  (The 'find' manpage and info page have complete listings.)

   1 find /etc -exec grep '[0-9][0-9]*[.][0-9][0-9]*[.][0-9][0-9]*[.][0-9][0-9]*' {} \;
   2 
   3 # Finds all IP addresses (xxx.xxx.xxx.xxx) in /etc directory files.
   4 # There a few extraneous hits. Can they be filtered out?
   5 
   6 # Possibly by:
   7 
   8 find /etc -type f -exec cat '{}' \; | tr -c '.[:digit:]' '\n' \
   9 | grep '^[^.][^.]*\.[^.][^.]*\.[^.][^.]*\.[^.][^.]*$'
  10 #
  11 #  [:digit:] is one of the character classes
  12 #+ introduced with the POSIX 1003.2 standard. 
  13 
  14 # Thanks, Stéphane Chazelas. 

Note

The -exec option to find should not be confused with the exec shell builtin.


Example 15-3. Badname, eliminate file names in current directory containing bad characters and whitespace.

   1 #!/bin/bash
   2 # badname.sh
   3 # Delete filenames in current directory containing bad characters.
   4 
   5 for filename in *
   6 do
   7   badname=`echo "$filename" | sed -n /[\+\{\;\"\\\=\?~\(\)\<\>\&\*\|\$]/p`
   8 # badname=`echo "$filename" | sed -n '/[+{;"\=?~()<>&*|$]/p'`  also works.
   9 # Deletes files containing these nasties:     + { ; " \ = ? ~ ( ) < > & * | $
  10 #
  11   rm $badname 2>/dev/null
  12 #             ^^^^^^^^^^^ Error messages deep-sixed.
  13 done
  14 
  15 # Now, take care of files containing all manner of whitespace.
  16 find . -name "* *" -exec rm -f {} \;
  17 # The path name of the file that _find_ finds replaces the "{}".
  18 # The '\' ensures that the ';' is interpreted literally, as end of command.
  19 
  20 exit 0
  21 
  22 #---------------------------------------------------------------------
  23 # Commands below this line will not execute because of _exit_ command.
  24 
  25 # An alternative to the above script:
  26 find . -name '*[+{;"\\=?~()<>&*|$ ]*' -maxdepth 0 \
  27 -exec rm -f '{}' \;
  28 #  The "-maxdepth 0" option ensures that _find_ will not search
  29 #+ subdirectories below $PWD.
  30 
  31 # (Thanks, S.C.)


Example 15-4. Deleting a file by its inode number

   1 #!/bin/bash
   2 # idelete.sh: Deleting a file by its inode number.
   3 
   4 #  This is useful when a filename starts with an illegal character,
   5 #+ such as ? or -.
   6 
   7 ARGCOUNT=1                      # Filename arg must be passed to script.
   8 E_WRONGARGS=70
   9 E_FILE_NOT_EXIST=71
  10 E_CHANGED_MIND=72
  11 
  12 if [ $# -ne "$ARGCOUNT" ]
  13 then
  14   echo "Usage: `basename $0` filename"
  15   exit $E_WRONGARGS
  16 fi  
  17 
  18 if [ ! -e "$1" ]
  19 then
  20   echo "File \""$1"\" does not exist."
  21   exit $E_FILE_NOT_EXIST
  22 fi  
  23 
  24 inum=`ls -i | grep "$1" | awk '{print $1}'`
  25 # inum = inode (index node) number of file
  26 # -----------------------------------------------------------------------
  27 # Every file has an inode, a record that holds its physical address info.
  28 # -----------------------------------------------------------------------
  29 
  30 echo; echo -n "Are you absolutely sure you want to delete \"$1\" (y/n)? "
  31 # The '-v' option to 'rm' also asks this.
  32 read answer
  33 case "$answer" in
  34 [nN]) echo "Changed your mind, huh?"
  35       exit $E_CHANGED_MIND
  36       ;;
  37 *)    echo "Deleting file \"$1\".";;
  38 esac
  39 
  40 find . -inum $inum -exec rm {} \;
  41 #                           ^^
  42 #        Curly brackets are placeholder
  43 #+       for text output by "find."
  44 echo "File "\"$1"\" deleted!"
  45 
  46 exit 0

The find command also works without the -exec option.

   1 #!/bin/bash
   2 #  Find suid root files.
   3 #  A strange suid file might indicate a security hole,
   4 #+ or even a system intrusion.
   5 
   6 directory="/usr/sbin"
   7 # Might also try /sbin, /bin, /usr/bin, /usr/local/bin, etc.
   8 permissions="+4000"  # suid root (dangerous!)
   9 
  10 
  11 for file in $( find "$directory" -perm "$permissions" )
  12 do
  13   ls -ltF --author "$file"
  14 done

See Example 15-30, Example 3-4, and Example 10-9 for scripts using find. Its manpage provides more detail on this complex and powerful command.

xargs

A filter for feeding arguments to a command, and also a tool for assembling the commands themselves. It breaks a data stream into small enough chunks for filters and commands to process. Consider it as a powerful replacement for backquotes. In situations where command substitution fails with a too many arguments error, substituting xargs often works. [1] Normally, xargs reads from stdin or from a pipe, but it can also be given the output of a file.

The default command for xargs is echo. This means that input piped to xargs may have linefeeds and other whitespace characters stripped out.

 bash$ ls -l
 total 0
 -rw-rw-r--    1 bozo  bozo         0 Jan 29 23:58 file1
 -rw-rw-r--    1 bozo  bozo         0 Jan 29 23:58 file2
 
 
 
 bash$ ls -l | xargs
 total 0 -rw-rw-r-- 1 bozo bozo 0 Jan 29 23:58 file1 -rw-rw-r-- 1 bozo bozo 0 Jan...
 
 
 
 bash$ find ~/mail -type f | xargs grep "Linux"
 ./misc:User-Agent: slrn/0.9.8.1 (Linux)
 ./sent-mail-jul-2005: hosted by the Linux Documentation Project.
 ./sent-mail-jul-2005: (Linux Documentation Project Site, rtf version)
 ./sent-mail-jul-2005: Subject: Criticism of Bozo's Windows/Linux article
 ./sent-mail-jul-2005: while mentioning that the Linux ext2/ext3 filesystem
 . . .
 	      

ls | xargs -p -l gzip gzips every file in current directory, one at a time, prompting before each operation.

Note

Note that xargs processes the arguments passed to it sequentially, one at a time.

 bash$ find /usr/bin | xargs file
 /usr/bin:          directory
 /usr/bin/foomatic-ppd-options:          perl script text executable
 . . .
 	      

Tip

An interesting xargs option is -n NN, which limits to NN the number of arguments passed.

ls | xargs -n 8 echo lists the files in the current directory in 8 columns.

Tip

Another useful option is -0, in combination with find -print0 or grep -lZ. This allows handling arguments containing whitespace or quotes.

find / -type f -print0 | xargs -0 grep -liwZ GUI | xargs -0 rm -f

grep -rliwZ GUI / | xargs -0 rm -f

Either of the above will remove any file containing "GUI". (Thanks, S.C.)

Or:
   1 cat /proc/"$pid"/"$OPTION" | xargs -0 echo
   2 #  Formats output:         ^^^^^^^^^^^^^^^
   3 #  From Han Holl's fixup of "get-commandline.sh"
   4 #+ script in "/dev and /proc" chapter.


Example 15-5. Logfile: Using xargs to monitor system log

   1 #!/bin/bash
   2 
   3 # Generates a log file in current directory
   4 # from the tail end of /var/log/messages.
   5 
   6 # Note: /var/log/messages must be world readable
   7 # if this script invoked by an ordinary user.
   8 #         #root chmod 644 /var/log/messages
   9 
  10 LINES=5
  11 
  12 ( date; uname -a ) >>logfile
  13 # Time and machine name
  14 echo ---------------------------------------------------------- >>logfile
  15 tail -n $LINES /var/log/messages | xargs | fmt -s >>logfile
  16 echo >>logfile
  17 echo >>logfile
  18 
  19 exit 0
  20 
  21 #  Note:
  22 #  ----
  23 #  As Frank Wang points out,
  24 #+ unmatched quotes (either single or double quotes) in the source file
  25 #+ may give xargs indigestion.
  26 #
  27 #  He suggests the following substitution for line 15:
  28 #  tail -n $LINES /var/log/messages | tr -d "\"'" | xargs | fmt -s >>logfile
  29 
  30 
  31 
  32 #  Exercise:
  33 #  --------
  34 #  Modify this script to track changes in /var/log/messages at intervals
  35 #+ of 20 minutes.
  36 #  Hint: Use the "watch" command. 

As in find, a curly bracket pair serves as a placeholder for replacement text.


Example 15-6. Copying files in current directory to another

   1 #!/bin/bash
   2 # copydir.sh
   3 
   4 #  Copy (verbose) all files in current directory ($PWD)
   5 #+ to directory specified on command line.
   6 
   7 E_NOARGS=65
   8 
   9 if [ -z "$1" ]   # Exit if no argument given.
  10 then
  11   echo "Usage: `basename $0` directory-to-copy-to"
  12   exit $E_NOARGS
  13 fi  
  14 
  15 ls . | xargs -i -t cp ./{} $1
  16 #            ^^ ^^      ^^
  17 #  -t is "verbose" (output command line to stderr) option.
  18 #  -i is "replace strings" option.
  19 #  {} is a placeholder for output text.
  20 #  This is similar to the use of a curly bracket pair in "find."
  21 #
  22 #  List the files in current directory (ls .),
  23 #+ pass the output of "ls" as arguments to "xargs" (-i -t options),
  24 #+ then copy (cp) these arguments ({}) to new directory ($1).  
  25 #
  26 #  The net result is the exact equivalent of
  27 #+   cp * $1
  28 #+ unless any of the filenames has embedded "whitespace" characters.
  29 
  30 exit 0


Example 15-7. Killing processes by name

   1 #!/bin/bash
   2 # kill-byname.sh: Killing processes by name.
   3 # Compare this script with kill-process.sh.
   4 
   5 #  For instance,
   6 #+ try "./kill-byname.sh xterm" --
   7 #+ and watch all the xterms on your desktop disappear.
   8 
   9 #  Warning:
  10 #  -------
  11 #  This is a fairly dangerous script.
  12 #  Running it carelessly (especially as root)
  13 #+ can cause data loss and other undesirable effects.
  14 
  15 E_BADARGS=66
  16 
  17 if test -z "$1"  # No command line arg supplied?
  18 then
  19   echo "Usage: `basename $0` Process(es)_to_kill"
  20   exit $E_BADARGS
  21 fi
  22 
  23 
  24 PROCESS_NAME="$1"
  25 ps ax | grep "$PROCESS_NAME" | awk '{print $1}' | xargs -i kill {} 2&>/dev/null
  26 #                                                       ^^      ^^
  27 
  28 # ---------------------------------------------------------------
  29 # Notes:
  30 # -i is the "replace strings" option to xargs.
  31 # The curly brackets are the placeholder for the replacement.
  32 # 2&>/dev/null suppresses unwanted error messages.
  33 #
  34 # Can  grep "$PROCESS_NAME" be replaced by pidof "$PROCESS_NAME"?
  35 # ---------------------------------------------------------------
  36 
  37 exit $?
  38 
  39 #  The "killall" command has the same effect as this script,
  40 #+ but using it is not quite as educational.


Example 15-8. Word frequency analysis using xargs

   1 #!/bin/bash
   2 # wf2.sh: Crude word frequency analysis on a text file.
   3 
   4 # Uses 'xargs' to decompose lines of text into single words.
   5 # Compare this example to the "wf.sh" script later on.
   6 
   7 
   8 # Check for input file on command line.
   9 ARGS=1
  10 E_BADARGS=65
  11 E_NOFILE=66
  12 
  13 if [ $# -ne "$ARGS" ]
  14 # Correct number of arguments passed to script?
  15 then
  16   echo "Usage: `basename $0` filename"
  17   exit $E_BADARGS
  18 fi
  19 
  20 if [ ! -f "$1" ]       # Check if file exists.
  21 then
  22   echo "File \"$1\" does not exist."
  23   exit $E_NOFILE
  24 fi
  25 
  26 
  27 
  28 ########################################################
  29 cat "$1" | xargs -n1 | \
  30 #  List the file, one word per line. 
  31 tr A-Z a-z | \
  32 #  Shift characters to lowercase.
  33 sed -e 's/\.//g'  -e 's/\,//g' -e 's/ /\
  34 /g' | \
  35 #  Filter out periods and commas, and
  36 #+ change space between words to linefeed,
  37 sort | uniq -c | sort -nr
  38 #  Finally prefix occurrence count and sort numerically.
  39 ########################################################
  40 
  41 #  This does the same job as the "wf.sh" example,
  42 #+ but a bit more ponderously, and it runs more slowly (why?).
  43 
  44 exit 0

expr

All-purpose expression evaluator: Concatenates and evaluates the arguments according to the operation given (arguments must be separated by spaces). Operations may be arithmetic, comparison, string, or logical.

expr 3 + 5

returns 8

expr 5 % 3

returns 2

expr 1 / 0

returns the error message, expr: division by zero

Illegal arithmetic operations not allowed.

expr 5 \* 3

returns 15

The multiplication operator must be escaped when used in an arithmetic expression with expr.

y=`expr $y + 1`

Increment a variable, with the same effect as let y=y+1 and y=$(($y+1)). This is an example of arithmetic expansion.

z=`expr substr $string $position $length`

Extract substring of $length characters, starting at $position.


Example 15-9. Using expr

   1 #!/bin/bash
   2 
   3 # Demonstrating some of the uses of 'expr'
   4 # =======================================
   5 
   6 echo
   7 
   8 # Arithmetic Operators
   9 # ---------- ---------
  10 
  11 echo "Arithmetic Operators"
  12 echo
  13 a=`expr 5 + 3`
  14 echo "5 + 3 = $a"
  15 
  16 a=`expr $a + 1`
  17 echo
  18 echo "a + 1 = $a"
  19 echo "(incrementing a variable)"
  20 
  21 a=`expr 5 % 3`
  22 # modulo
  23 echo
  24 echo "5 mod 3 = $a"
  25 
  26 echo
  27 echo
  28 
  29 # Logical Operators
  30 # ------- ---------
  31 
  32 #  Returns 1 if true, 0 if false,
  33 #+ opposite of normal Bash convention.
  34 
  35 echo "Logical Operators"
  36 echo
  37 
  38 x=24
  39 y=25
  40 b=`expr $x = $y`         # Test equality.
  41 echo "b = $b"            # 0  ( $x -ne $y )
  42 echo
  43 
  44 a=3
  45 b=`expr $a \> 10`
  46 echo 'b=`expr $a \> 10`, therefore...'
  47 echo "If a > 10, b = 0 (false)"
  48 echo "b = $b"            # 0  ( 3 ! -gt 10 )
  49 echo
  50 
  51 b=`expr $a \< 10`
  52 echo "If a < 10, b = 1 (true)"
  53 echo "b = $b"            # 1  ( 3 -lt 10 )
  54 echo
  55 # Note escaping of operators.
  56 
  57 b=`expr $a \<= 3`
  58 echo "If a <= 3, b = 1 (true)"
  59 echo "b = $b"            # 1  ( 3 -le 3 )
  60 # There is also a "\>=" operator (greater than or equal to).
  61 
  62 
  63 echo
  64 echo
  65 
  66 
  67 
  68 # String Operators
  69 # ------ ---------
  70 
  71 echo "String Operators"
  72 echo
  73 
  74 a=1234zipper43231
  75 echo "The string being operated upon is \"$a\"."
  76 
  77 # length: length of string
  78 b=`expr length $a`
  79 echo "Length of \"$a\" is $b."
  80 
  81 # index: position of first character in substring
  82 #        that matches a character in string
  83 b=`expr index $a 23`
  84 echo "Numerical position of first \"2\" in \"$a\" is \"$b\"."
  85 
  86 # substr: extract substring, starting position & length specified
  87 b=`expr substr $a 2 6`
  88 echo "Substring of \"$a\", starting at position 2,\
  89 and 6 chars long is \"$b\"."
  90 
  91 
  92 #  The default behavior of the 'match' operations is to
  93 #+ search for the specified match at the ***beginning*** of the string.
  94 #
  95 #        uses Regular Expressions
  96 b=`expr match "$a" '[0-9]*'`               #  Numerical count.
  97 echo Number of digits at the beginning of \"$a\" is $b.
  98 b=`expr match "$a" '\([0-9]*\)'`           #  Note that escaped parentheses
  99 #                   ==      ==              + trigger substring match.
 100 echo "The digits at the beginning of \"$a\" are \"$b\"."
 101 
 102 echo
 103 
 104 exit 0

Important

The : operator can substitute for match. For example, b=`expr $a : [0-9]*` is the exact equivalent of b=`expr match $a [0-9]*` in the above listing.

   1 #!/bin/bash
   2 
   3 echo
   4 echo "String operations using \"expr \$string : \" construct"
   5 echo "==================================================="
   6 echo
   7 
   8 a=1234zipper5FLIPPER43231
   9 
  10 echo "The string being operated upon is \"`expr "$a" : '\(.*\)'`\"."
  11 #     Escaped parentheses grouping operator.            ==  ==
  12 
  13 #       ***************************
  14 #+          Escaped parentheses
  15 #+           match a substring
  16 #       ***************************
  17 
  18 
  19 #  If no escaped parentheses...
  20 #+ then 'expr' converts the string operand to an integer.
  21 
  22 echo "Length of \"$a\" is `expr "$a" : '.*'`."   # Length of string
  23 
  24 echo "Number of digits at the beginning of \"$a\" is `expr "$a" : '[0-9]*'`."
  25 
  26 # ------------------------------------------------------------------------- #
  27 
  28 echo
  29 
  30 echo "The digits at the beginning of \"$a\" are `expr "$a" : '\([0-9]*\)'`."
  31 #                                                             ==      ==
  32 echo "The first 7 characters of \"$a\" are `expr "$a" : '\(.......\)'`."
  33 #         =====                                          ==       ==
  34 # Again, escaped parentheses force a substring match.
  35 #
  36 echo "The last 7 characters of \"$a\" are `expr "$a" : '.*\(.......\)'`."
  37 #         ====                  end of string operator  ^^
  38 #  (actually means skip over one or more of any characters until specified
  39 #+  substring)
  40 
  41 echo
  42 
  43 exit 0

The above script illustrates how expr uses the escaped parentheses -- \( ... \) -- grouping operator in tandem with regular expression parsing to match a substring. Here is a another example, this time from "real life."
   1 # Strip the whitespace from the beginning and end.
   2 LRFDATE=`expr "$LRFDATE" : '[[:space:]]*\(.*\)[[:space:]]*$'`
   3 
   4 #  From Peter Knowles' "booklistgen.sh" script
   5 #+ for converting files to Sony Librie/PRS-50X format.
   6 #  (http://booklistgensh.peterknowles.com)

Perl, sed, and awk have far superior string parsing facilities. A short sed or awk "subroutine" within a script (see Section 33.3) is an attractive alternative to expr.

See Section 9.2 for more on using expr in string operations.

Notes

[1]

And even when xargs is not strictly necessary, it can speed up execution of a command involving batch-processing of multiple files.