Chapter 18. Here Documents
Here and now, boys. | |
| Aldous Huxley, Island |
A here document is a special-purpose code block. It uses a form of I/O redirection to feed a command list to an interactive program or a command, such as ftp, cat, or the ex text editor.
1 COMMAND <<InputComesFromHERE 2 ... 3 InputComesFromHERE |
A limit string delineates (frames) the command list. The special symbol << designates the limit string. This has the effect of redirecting the output of a file into the stdin of the program or command. It is similar to interactive-program < command-file, where command-file contains
1 command #1 2 command #2 3 ... |
The here document alternative looks like this:
1 #!/bin/bash 2 interactive-program <<LimitString 3 command #1 4 command #2 5 ... 6 LimitString |
Choose a limit string sufficiently unusual that it will not occur anywhere in the command list and confuse matters.
Note that here documents may sometimes be used to good effect with non-interactive utilities and commands, such as, for example, wall.
Example 18-1. broadcast: Sends message to everyone logged in
1 #!/bin/bash 2 3 wall <<zzz23EndOfMessagezzz23 4 E-mail your noontime orders for pizza to the system administrator. 5 (Add an extra dollar for anchovy or mushroom topping.) 6 # Additional message text goes here. 7 # Note: 'wall' prints comment lines. 8 zzz23EndOfMessagezzz23 9 10 # Could have been done more efficiently by 11 # wall <message-file 12 # However, embedding the message template in a script 13 #+ is a quick-and-dirty one-off solution. 14 15 exit 0 |
Even such unlikely candidates as vi lend themselves to here documents.
Example 18-2. dummyfile: Creates a 2-line dummy file
1 #!/bin/bash 2 3 # Non-interactive use of 'vi' to edit a file. 4 # Emulates 'sed'. 5 6 E_BADARGS=65 7 8 if [ -z "$1" ] 9 then 10 echo "Usage: `basename $0` filename" 11 exit $E_BADARGS 12 fi 13 14 TARGETFILE=$1 15 16 # Insert 2 lines in file, then save. 17 #--------Begin here document-----------# 18 vi $TARGETFILE <<x23LimitStringx23 19 i 20 This is line 1 of the example file. 21 This is line 2 of the example file. 22 ^[ 23 ZZ 24 x23LimitStringx23 25 #----------End here document-----------# 26 27 # Note that ^[ above is a literal escape 28 #+ typed by Control-V <Esc>. 29 30 # Bram Moolenaar points out that this may not work with 'vim', 31 #+ because of possible problems with terminal interaction. 32 33 exit 0 |
The above script could just as effectively have been implemented with ex, rather than vi. Here documents containing a list of ex commands are common enough to form their own category, known as ex scripts.
1 #!/bin/bash 2 # Replace all instances of "Smith" with "Jones" 3 #+ in files with a ".txt" filename suffix. 4 5 ORIGINAL=Smith 6 REPLACEMENT=Jones 7 8 for word in $(fgrep -l $ORIGINAL *.txt) 9 do 10 # ------------------------------------- 11 ex $word <<EOF 12 :%s/$ORIGINAL/$REPLACEMENT/g 13 :wq 14 EOF 15 # :%s is the "ex" substitution command. 16 # :wq is write-and-quit. 17 # ------------------------------------- 18 done |
Analogous to "ex scripts" are cat scripts.
Example 18-3. Multi-line message using cat
1 #!/bin/bash 2 3 # 'echo' is fine for printing single line messages, 4 #+ but somewhat problematic for for message blocks. 5 # A 'cat' here document overcomes this limitation. 6 7 cat <<End-of-message 8 ------------------------------------- 9 This is line 1 of the message. 10 This is line 2 of the message. 11 This is line 3 of the message. 12 This is line 4 of the message. 13 This is the last line of the message. 14 ------------------------------------- 15 End-of-message 16 17 # Replacing line 7, above, with 18 #+ cat > $Newfile <<End-of-message 19 #+ ^^^^^^^^^^ 20 #+ writes the output to the file $Newfile, rather than to stdout. 21 22 exit 0 23 24 25 #-------------------------------------------- 26 # Code below disabled, due to "exit 0" above. 27 28 # S.C. points out that the following also works. 29 echo "------------------------------------- 30 This is line 1 of the message. 31 This is line 2 of the message. 32 This is line 3 of the message. 33 This is line 4 of the message. 34 This is the last line of the message. 35 -------------------------------------" 36 # However, text may not include double quotes unless they are escaped. |
The - option to mark a here document limit string (<<-LimitString) suppresses leading tabs (but not spaces) in the output. This may be useful in making a script more readable.
Example 18-4. Multi-line message, with tabs suppressed
1 #!/bin/bash 2 # Same as previous example, but... 3 4 # The - option to a here document <<- 5 #+ suppresses leading tabs in the body of the document, 6 #+ but *not* spaces. 7 8 cat <<-ENDOFMESSAGE 9 This is line 1 of the message. 10 This is line 2 of the message. 11 This is line 3 of the message. 12 This is line 4 of the message. 13 This is the last line of the message. 14 ENDOFMESSAGE 15 # The output of the script will be flush left. 16 # Leading tab in each line will not show. 17 18 # Above 5 lines of "message" prefaced by a tab, not spaces. 19 # Spaces not affected by <<- . 20 21 # Note that this option has no effect on *embedded* tabs. 22 23 exit 0 |
A here document supports parameter and command substitution. It is therefore possible to pass different parameters to the body of the here document, changing its output accordingly.
Example 18-5. Here document with parameter substitution
1 #!/bin/bash 2 # Another 'cat' here document, using parameter substitution. 3 4 # Try it with no command line parameters, ./scriptname 5 # Try it with one command line parameter, ./scriptname Mortimer 6 # Try it with one two-word quoted command line parameter, 7 # ./scriptname "Mortimer Jones" 8 9 CMDLINEPARAM=1 # Expect at least command line parameter. 10 11 if [ $# -ge $CMDLINEPARAM ] 12 then 13 NAME=$1 # If more than one command line param, 14 #+ then just take the first. 15 else 16 NAME="John Doe" # Default, if no command line parameter. 17 fi 18 19 RESPONDENT="the author of this fine script" 20 21 22 cat <<Endofmessage 23 24 Hello, there, $NAME. 25 Greetings to you, $NAME, from $RESPONDENT. 26 27 # This comment shows up in the output (why?). 28 29 Endofmessage 30 31 # Note that the blank lines show up in the output. 32 # So does the "comment". 33 34 exit 0 |
This is a useful script containing a here document with parameter substitution.
Example 18-6. Upload a file pair to Sunsite incoming directory
1 #!/bin/bash 2 # upload.sh 3 4 # Upload file pair (Filename.lsm, Filename.tar.gz) 5 #+ to incoming directory at Sunsite/UNC (ibiblio.org). 6 # Filename.tar.gz is the tarball itself. 7 # Filename.lsm is the descriptor file. 8 # Sunsite requires "lsm" file, otherwise will bounce contributions. 9 10 11 E_ARGERROR=65 12 13 if [ -z "$1" ] 14 then 15 echo "Usage: `basename $0` Filename-to-upload" 16 exit $E_ARGERROR 17 fi 18 19 20 Filename=`basename $1` # Strips pathname out of file name. 21 22 Server="ibiblio.org" 23 Directory="/incoming/Linux" 24 # These need not be hard-coded into script, 25 #+ but may instead be changed to command line argument. 26 27 Password="your.e-mail.address" # Change above to suit. 28 29 ftp -n $Server <<End-Of-Session 30 # -n option disables auto-logon 31 32 user anonymous "$Password" 33 binary 34 bell # Ring 'bell' after each file transfer. 35 cd $Directory 36 put "$Filename.lsm" 37 put "$Filename.tar.gz" 38 bye 39 End-Of-Session 40 41 exit 0 |
Quoting or escaping the "limit string" at the head of a here document disables parameter substitution within its body.
Example 18-7. Parameter substitution turned off
1 #!/bin/bash 2 # A 'cat' here-document, but with parameter substitution disabled. 3 4 NAME="John Doe" 5 RESPONDENT="the author of this fine script" 6 7 cat <<'Endofmessage' 8 9 Hello, there, $NAME. 10 Greetings to you, $NAME, from $RESPONDENT. 11 12 Endofmessage 13 14 # No parameter substitution when the "limit string" is quoted or escaped. 15 # Either of the following at the head of the here document would have 16 #+ the same effect. 17 # cat <<"Endofmessage" 18 # cat <<\Endofmessage 19 20 exit 0 |
Disabling parameter substitution permits outputting literal text. Generating scripts or even program code is one use for this.
Example 18-8. A script that generates another script
1 #!/bin/bash 2 # generate-script.sh 3 # Based on an idea by Albert Reiner. 4 5 OUTFILE=generated.sh # Name of the file to generate. 6 7 8 # ----------------------------------------------------------- 9 # 'Here document containing the body of the generated script. 10 ( 11 cat <<'EOF' 12 #!/bin/bash 13 14 echo "This is a generated shell script." 15 # Note that since we are inside a subshell, 16 #+ we can't access variables in the "outside" script. 17 18 echo "Generated file will be named: $OUTFILE" 19 # Above line will not work as normally expected 20 #+ because parameter expansion has been disabled. 21 # Instead, the result is literal output. 22 23 a=7 24 b=3 25 26 let "c = $a * $b" 27 echo "c = $c" 28 29 exit 0 30 EOF 31 ) > $OUTFILE 32 # ----------------------------------------------------------- 33 34 # Quoting the 'limit string' prevents variable expansion 35 #+ within the body of the above 'here document.' 36 # This permits outputting literal strings in the output file. 37 38 if [ -f "$OUTFILE" ] 39 then 40 chmod 755 $OUTFILE 41 # Make the generated file executable. 42 else 43 echo "Problem in creating file: \"$OUTFILE\"" 44 fi 45 46 # This method can also be used for generating 47 #+ C programs, Perl programs, Python programs, Makefiles, 48 #+ and the like. 49 50 exit 0 |
It is possible to set a variable from the output of a here document.
1 variable=$(cat <<SETVAR 2 This variable 3 runs over multiple lines. 4 SETVAR) 5 6 echo "$variable" |
A here document can supply input to a function in the same script.
Example 18-9. Here documents and functions
1 #!/bin/bash
2 # here-function.sh
3
4 GetPersonalData ()
5 {
6 read firstname
7 read lastname
8 read address
9 read city
10 read state
11 read zipcode
12 } # This certainly looks like an interactive function, but...
13
14
15 # Supply input to the above function.
16 GetPersonalData <<RECORD001
17 Bozo
18 Bozeman
19 2726 Nondescript Dr.
20 Baltimore
21 MD
22 21226
23 RECORD001
24
25
26 echo
27 echo "$firstname $lastname"
28 echo "$address"
29 echo "$city, $state $zipcode"
30 echo
31
32 exit 0 |
It is possible to use : as a dummy command accepting output from a here document. This, in effect, creates an "anonymous" here document.
Example 18-10. "Anonymous" Here Document
1 #!/bin/bash
2
3 : <<TESTVARIABLES
4 ${HOSTNAME?}${USER?}${MAIL?} # Print error message if one of the variables not set.
5 TESTVARIABLES
6
7 exit 0 |
 | A variation of the above technique permits "commenting out" blocks of code. |
Example 18-11. Commenting out a block of code
1 #!/bin/bash
2 # commentblock.sh
3
4 : <<COMMENTBLOCK
5 echo "This line will not echo."
6 This is a comment line missing the "#" prefix.
7 This is another comment line missing the "#" prefix.
8
9 &*@!!++=
10 The above line will cause no error message,
11 because the Bash interpreter will ignore it.
12 COMMENTBLOCK
13
14 echo "Exit value of above \"COMMENTBLOCK\" is $?." # 0
15 # No error shown.
16 echo
17
18
19 # The above technique also comes in useful for commenting out
20 #+ a block of working code for debugging purposes.
21 # This saves having to put a "#" at the beginning of each line,
22 #+ then having to go back and delete each "#" later.
23
24 echo "Just before commented-out code block."
25 # The lines of code between the double-dashed lines will not execute.
26 # ===================================================================
27 : <<DEBUGXXX
28 for file in *
29 do
30 cat "$file"
31 done
32 DEBUGXXX
33 # ===================================================================
34 echo "Just after commented-out code block."
35
36 exit 0
37
38
39
40 ######################################################################
41 # Note, however, that if a bracketed variable is contained within
42 #+ the commented-out code block,
43 #+ then this could cause problems.
44 # for example:
45
46
47 #/!/bin/bash
48
49 : <<COMMENTBLOCK
50 echo "This line will not echo."
51 &*@!!++=
52 ${foo_bar_bazz?}
53 $(rm -rf /tmp/foobar/)
54 $(touch my_build_directory/cups/Makefile)
55 COMMENTBLOCK
56
57
58 $ sh commented-bad.sh
59 commented-bad.sh: line 3: foo_bar_bazz: parameter null or not set
60
61 # The remedy for this is to strong-quote the 'COMMENTBLOCK' in line 49, above.
62
63 : <<'COMMENTBLOCK'
64
65 # Thank you, Kurt Pfeifle, for pointing this out. |
| Yet another twist of this nifty trick makes "self-documenting" scripts possible. |
Example 18-12. A self-documenting script
1 #!/bin/bash 2 # self-document.sh: self-documenting script 3 # Modification of "colm.sh". 4 5 DOC_REQUEST=70 6 7 if [ "$1" = "-h" -o "$1" = "--help" ] # Request help. 8 then 9 echo; echo "Usage: $0 [directory-name]"; echo 10 sed --silent -e '/DOCUMENTATIONXX$/,/^DOCUMENTATIONXX$/p' "$0" | 11 sed -e '/DOCUMENTATIONXX$/d'; exit $DOC_REQUEST; fi 12 13 14 : <<DOCUMENTATIONXX 15 List the statistics of a specified directory in tabular format. 16 --------------------------------------------------------------- 17 The command line parameter gives the directory to be listed. 18 If no directory specified or directory specified cannot be read, 19 then list the current working directory. 20 21 DOCUMENTATIONXX 22 23 if [ -z "$1" -o ! -r "$1" ] 24 then 25 directory=. 26 else 27 directory="$1" 28 fi 29 30 echo "Listing of "$directory":"; echo 31 (printf "PERMISSIONS LINKS OWNER GROUP SIZE MONTH DAY HH:MM PROG-NAME\n" \ 32 ; ls -l "$directory" | sed 1d) | column -t 33 34 exit 0 |
Using a cat script is an alternate way of accomplishing this.
1 DOC_REQUEST=70 2 3 if [ "$1" = "-h" -o "$1" = "--help" ] # Request help. 4 then # Use a "cat script" . . . 5 cat <<DOCUMENTATIONXX 6 List the statistics of a specified directory in tabular format. 7 --------------------------------------------------------------- 8 The command line parameter gives the directory to be listed. 9 If no directory specified or directory specified cannot be read, 10 then list the current working directory. 11 12 DOCUMENTATIONXX 13 exit $DOC_REQUEST 14 fi |
See also Example A-29 for one more excellent example of a self-documenting script.
 | Here documents create temporary files, but these files are deleted after opening and are not accessible to any other process.
|
 | Some utilities will not work inside a here document. |
 | The closing limit string, on the final line of a here document, must start in the first character position. There can be no leading whitespace. Trailing whitespace after the limit string likewise causes unexpected behavior. The whitespace prevents the limit string from being recognized.
|
For those tasks too complex for a "here document", consider using the expect scripting language, which is specifically tailored for feeding input into interactive programs.
18.1. Here Strings
A here string can be considered as a stripped-down form of a here document. It consists of nothing more than COMMAND <<<$WORD, where $WORD is expanded and fed to the stdin of COMMAND.
As a simple example, consider this alternative to the echo-grep construction.
1 # Instead of: 2 if echo "$VAR" | grep -q txt # if [[ $VAR = *txt* ]] 3 # etc. 4 5 # Try: 6 if grep -q "txt" <<< "$VAR" 7 then 8 echo "$VAR contains the substring sequence \"txt\"" 9 fi 10 # Thank you, Sebastian Kaminski, for the suggestion. |
Or, in combination with read:
1 String="This is a string of words."
2
3 read -r -a Words <<< "$String"
4 # The -a option to "read"
5 #+ assigns the resulting values to successive members of an array.
6
7 echo "First word in String is: ${Words[0]}" # This
8 echo "Second word in String is: ${Words[1]}" # is
9 echo "Third word in String is: ${Words[2]}" # a
10 echo "Fourth word in String is: ${Words[3]}" # string
11 echo "Fifth word in String is: ${Words[4]}" # of
12 echo "Sixth word in String is: ${Words[5]}" # words.
13 echo "Seventh word in String is: ${Words[6]}" # (null)
14 # Past end of $String.
15
16 # Thank you, Francisco Lobo, for the suggestion. |
Example 18-13. Prepending a line to a file
1 #!/bin/bash 2 # prepend.sh: Add text at beginning of file. 3 # 4 # Example contributed by Kenny Stauffer, 5 #+ and slightly modified by document author. 6 7 8 E_NOSUCHFILE=65 9 10 read -p "File: " file # -p arg to 'read' displays prompt. 11 if [ ! -e "$file" ] 12 then # Bail out if no such file. 13 echo "File $file not found." 14 exit $E_NOSUCHFILE 15 fi 16 17 read -p "Title: " title 18 cat - $file <<<$title > $file.new 19 20 echo "Modified file is $file.new" 21 22 exit 0 23 24 # from 'man bash': 25 # Here Strings 26 # A variant of here documents, the format is: 27 # 28 # <<<word 29 # 30 # The word is expanded and supplied to the command on its standard input. |
Example 18-14. Parsing a mailbox
1 #!/bin/bash
2 # Script by Francisco Lobo,
3 #+ and slightly modified and commented by ABS Guide author.
4 # Used in ABS Guide with permission. (Thank you!)
5
6 # This script will not run under Bash versions < 3.0.
7
8
9 E_MISSING_ARG=67
10 if [ -z "$1" ]
11 then
12 echo "Usage: $0 mailbox-file"
13 exit $E_MISSING_ARG
14 fi
15
16 mbox_grep() # Parse mailbox file.
17 {
18 declare -i body=0 match=0
19 declare -a date sender
20 declare mail header value
21
22
23 while IFS= read -r mail
24 # ^^^^ Reset $IFS.
25 # Otherwise "read" will strip leading & trailing space from its input.
26
27 do
28 if [[ $mail =~ "^From " ]] # Match "From" field in message.
29 then
30 (( body = 0 )) # "Zero out" variables.
31 (( match = 0 ))
32 unset date
33
34 elif (( body ))
35 then
36 (( match ))
37 # echo "$mail"
38 # Uncomment above line if you want entire body of message to display.
39
40 elif [[ $mail ]]; then
41 IFS=: read -r header value <<< "$mail"
42 # ^^^ "here string"
43
44 case "$header" in
45 [Ff][Rr][Oo][Mm] ) [[ $value =~ "$2" ]] && (( match++ )) ;;
46 # Match "From" line.
47 [Dd][Aa][Tt][Ee] ) read -r -a date <<< "$value" ;;
48 # ^^^
49 # Match "Date" line.
50 [Rr][Ee][Cc][Ee][Ii][Vv][Ee][Dd] ) read -r -a sender <<< "$value" ;;
51 # ^^^
52 # Match IP Address (may be spoofed).
53 esac
54
55 else
56 (( body++ ))
57 (( match )) &&
58 echo "MESSAGE ${date:+of: ${date[*]} }"
59 # Entire $date array ^
60 echo "IP address of sender: ${sender[1]}"
61 # Second field of "Received" line ^
62
63 fi
64
65
66 done < "$1" # Redirect stdout of file into loop.
67 }
68
69
70 mbox_grep "$1" # Send mailbox file to function.
71
72 exit $?
73
74 # Exercises:
75 # ---------
76 # 1) Break the single function, above, into multiple functions,
77 #+ for the sake of readability.
78 # 2) Add additional parsing to the script, checking for various keywords.
79
80
81
82 $ mailbox_grep.sh scam_mail
83 --> MESSAGE of Thu, 5 Jan 2006 08:00:56 -0500 (EST)
84 --> IP address of sender: 196.3.62.4 |
Exercise: Find other uses for here strings.