\n\nstringZ=abcABC123ABCabc\n\necho ${#stringZ} # 15\necho `expr length $stringZ` # 15\necho `expr \"$stringZ\" : '.*'` # 15<\/pre>\n<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/dd>\n<\/dl>\n<\/div>\n\n Example 10-1. Inserting a blank line between paragraphs in a text file<\/b><\/p>\n\n\n\n\n#!\/bin\/bash\n# paragraph-space.sh\n# Ver. 2.1, Reldate 29Jul12 [fixup]\n\n# Inserts a blank line between paragraphs of a single-spaced text file.\n# Usage: $0 <FILENAME\n\nMINLEN=60 # Change this value? It's a judgment call.\n# Assume lines shorter than $MINLEN characters ending in a period\n#+ terminate a paragraph. See exercises below.\n\nwhile read line # For as many lines as the input file has ...\ndo\n echo \"$line\" # Output the line itself.\n\n len=${#line}\n if [[ \"$len\" -lt \"$MINLEN\" && \"$line\" =~ [*{\\.}]$ ]]\n# if [[ \"$len\" -lt \"$MINLEN\" && \"$line\" =~ \\[*\\.\\] ]]\n# An update to Bash broke the previous version of this script. Ouch!\n# Thank you, Halim Srama, for pointing this out and suggesting a fix.\n then echo # Add a blank line immediately\n fi #+ after a short line terminated by a period.\ndone\n\nexit\n\n# Exercises:\n# ---------\n# 1) The script usually inserts a blank line at the end\n#+ of the target file. Fix this.\n# 2) Line 17 only considers periods as sentence terminators.\n# Modify this to include other common end-of-sentence characters,\n#+ such as ?, !, and \".<\/pre>\n<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/div>\n\n Length of Matching Substring at Beginning of String<\/b><\/p>\n\n- expr match “$string” ‘$substring’<\/dt>\n
- $substring<\/i><\/tt>\u00a0is a\u00a0regular expression.<\/dd>\n
- expr “$string” : ‘$substring’<\/dt>\n
- $substring<\/i><\/tt>\u00a0is a regular expression.<\/p>\n
\n\n\n\nstringZ=abcABC123ABCabc\n# |------|\n# 12345678\n\necho `expr match \"$stringZ\" 'abc[A-Z]*.2'` # 8\necho `expr \"$stringZ\" : 'abc[A-Z]*.2'` # 8<\/pre>\n<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/dd>\n<\/dl>\n<\/div>\n\n Index<\/b><\/p>\n\n- expr index $string $substring<\/dt>\n
- Numerical position in $string of first character in $substring that matches.<\/p>\n
\n\n\n\nstringZ=abcABC123ABCabc\n# 123456 ...\necho `expr index \"$stringZ\" C12` # 6\n # C position.\n\necho `expr index \"$stringZ\" 1c` # 3\n# 'c' (in #3 position) matches before '1'.<\/pre>\n<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\nThis is the near equivalent of\u00a0strchr()<\/i>\u00a0in\u00a0C<\/i>.<\/dd>\n<\/dl>\n<\/div>\n\n Substring Extraction<\/b><\/p>\n\n- ${string:position}<\/dt>\n
- Extracts substring from\u00a0$string<\/i><\/tt>\u00a0at\u00a0$position<\/i><\/tt>.If the\u00a0$string<\/tt>\u00a0parameter is\u00a0“*”\u00a0or\u00a0“@”, then this extracts the\u00a0positional parameters,\u00a0[1]\u00a0starting at\u00a0$position<\/tt>.<\/dd>\n
- ${string:position:length}<\/dt>\n
- Extracts\u00a0$length<\/i><\/tt>\u00a0characters of substring from\u00a0$string<\/i><\/tt>\u00a0at\u00a0$position<\/i><\/tt>.<\/p>\n
\n\n\n\nstringZ=abcABC123ABCabc\n# 0123456789.....\n# 0-based indexing.\n\necho ${stringZ:0} # abcABC123ABCabc\necho ${stringZ:1} # bcABC123ABCabc\necho ${stringZ:7} # 23ABCabc\n\necho ${stringZ:7:3} # 23A\n # Three characters of substring.\n\n# Is it possible to index from the right end of the string?\n\necho ${stringZ:-4} # abcABC123ABCabc\n# Defaults to full string, as in ${parameter:-default}.\n# However . . .\n\necho ${stringZ:(-4)} # Cabc \necho ${stringZ: -4} # Cabc\n# Now, it works.\n# Parentheses or added space \"escape\" the position parameter.\n\n# Thank you, Dan Jacobson, for pointing this out.<\/pre>\n<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\nThe\u00a0position<\/i>\u00a0and\u00a0length<\/i>\u00a0arguments can be\u00a0“parameterized,”\u00a0that is, represented as a variable, rather than as a numerical constant.<\/p>\n\n Example 10-2. Generating an 8-character\u00a0“random”\u00a0string<\/b><\/p>\n\n\n\n\n#!\/bin\/bash\n# rand-string.sh\n# Generating an 8-character \"random\" string.\n\nif [ -n \"$1\" ] # If command-line argument present,\nthen #+ then set start-string to it.\n str0=\"$1\"\nelse # Else use PID of script as start-string.\n str0=\"$\"\nfi\n\nPOS=2 # Starting from position 2 in the string.\nLEN=8 # Extract eight characters.\n\nstr1=$( echo \"$str0\" | md5sum | md5sum )\n# Doubly scramble ^^^^^^ ^^^^^^\n#+ by piping and repiping to md5sum.\n\nrandstring=\"${str1:$POS:$LEN}\"\n# Can parameterize ^^^^ ^^^^\n\necho \"$randstring\"\n\nexit $?\n\n# bozo$ .\/rand-string.sh my-password\n# 1bdd88c4\n\n# No, this is is not recommended\n#+ as a method of generating hack-proof passwords.<\/pre>\n<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/div>\nIf the\u00a0$string<\/tt>\u00a0parameter is\u00a0“*”\u00a0or\u00a0“@”, then this extracts a maximum of\u00a0$length<\/tt>\u00a0positional parameters, starting at\u00a0$position<\/tt>.<\/p>\n\n\n\n\necho ${*:2} # Echoes second and following positional parameters.\necho ${@:2} # Same as above.\n\necho ${*:2:3} # Echoes three positional parameters, starting at second.<\/pre>\n<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/dd>\n- expr substr $string $position $length<\/dt>\n
- Extracts\u00a0$length<\/i><\/tt>\u00a0characters from\u00a0$string<\/i><\/tt>\u00a0starting at\u00a0$position<\/i><\/tt>.<\/p>\n
\n\n\n\nstringZ=abcABC123ABCabc\n# 123456789......\n# 1-based indexing.\n\necho `expr substr $stringZ 1 2` # ab\necho `expr substr $stringZ 4 3` # ABC<\/pre>\n<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/dd>\n - expr match “$string” ‘\\($substring\\)’<\/dt>\n
- Extracts\u00a0$substring<\/i><\/tt>\u00a0at beginning of\u00a0$string<\/i><\/tt>, where\u00a0$substring<\/i><\/tt>\u00a0is a\u00a0regular expression.<\/dd>\n
- expr “$string” : ‘\\($substring\\)’<\/dt>\n
- Extracts\u00a0$substring<\/i><\/tt>\u00a0at beginning of\u00a0$string<\/i><\/tt>, where\u00a0$substring<\/i><\/tt>\u00a0is a regular expression.<\/p>\n
\n\n\n\nstringZ=abcABC123ABCabc\n# =======\t \n\necho `expr match \"$stringZ\" '\\(.[b-c]*[A-Z]..[0-9]\\)'` # abcABC1\necho `expr \"$stringZ\" : '\\(.[b-c]*[A-Z]..[0-9]\\)'` # abcABC1\necho `expr \"$stringZ\" : '\\(.......\\)'` # abcABC1\n# All of the above forms give an identical result.<\/pre>\n<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/dd>\n - expr match “$string” ‘.*\\($substring\\)’<\/dt>\n
- Extracts\u00a0$substring<\/i><\/tt>\u00a0at\u00a0end<\/em>\u00a0of\u00a0$string<\/i><\/tt>, where\u00a0$substring<\/i><\/tt>\u00a0is a regular expression.<\/dd>\n
- expr “$string” : ‘.*\\($substring\\)’<\/dt>\n
- Extracts\u00a0$substring<\/i><\/tt>\u00a0at\u00a0end<\/em>\u00a0of\u00a0$string<\/i><\/tt>, where\u00a0$substring<\/i><\/tt>\u00a0is a regular expression.<\/p>\n
\n\n\n\nstringZ=abcABC123ABCabc\n# ======\n\necho `expr match \"$stringZ\" '.*\\([A-C][A-C][A-C][a-c]*\\)'` # ABCabc\necho `expr \"$stringZ\" : '.*\\(......\\)'` # ABCabc<\/pre>\n<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/dd>\n<\/dl>\n<\/div>\n\n Substring Removal<\/b><\/p>\n\n- ${string#substring}<\/dt>\n
- Deletes shortest match of\u00a0$substring<\/i><\/tt>\u00a0from\u00a0front<\/em>\u00a0of\u00a0$string<\/i><\/tt>.<\/dd>\n
- ${string##substring}<\/dt>\n
- Deletes longest match of\u00a0$substring<\/i><\/tt>\u00a0from\u00a0front<\/em>\u00a0of\u00a0$string<\/i><\/tt>.<\/p>\n
\n\n\n\nstringZ=abcABC123ABCabc\n# |----| shortest\n# |----------| longest\n\necho ${stringZ#a*C} # 123ABCabc\n# Strip out shortest match between 'a' and 'C'.\n\necho ${stringZ##a*C} # abc\n# Strip out longest match between 'a' and 'C'.\n\n# You can parameterize the substrings.\n\nX='a*C'\n\necho ${stringZ#$X} # 123ABCabc\necho ${stringZ##$X} # abc\n # As above.<\/pre>\n<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/dd>\n- ${string%substring}<\/dt>\n
- Deletes shortest match of\u00a0$substring<\/i><\/tt>\u00a0from\u00a0back<\/em>\u00a0of\u00a0$string<\/i><\/tt>.For example:<\/p>\n
\n\n\n\n# Rename all filenames in $PWD with \"TXT\" suffix to a \"txt\" suffix.\n# For example, \"file1.TXT\" becomes \"file1.txt\" . . .\n\nSUFF=TXT\nsuff=txt\n\nfor i in $(ls *.$SUFF)\ndo\n mv -f $i ${i%.$SUFF}.$suff\n # Leave unchanged everything *except* the shortest pattern match\n #+ starting from the right-hand-side of the variable $i . . .\ndone ### This could be condensed into a \"one-liner\" if desired.\n\n# Thank you, Rory Winston.<\/pre>\n<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/dd>\n- ${string%%substring}<\/dt>\n
- Deletes longest match of\u00a0$substring<\/i><\/tt>\u00a0from\u00a0back<\/em>\u00a0of\u00a0$string<\/i><\/tt>.<\/p>\n
\n\n\n\nstringZ=abcABC123ABCabc\n# || shortest\n# |------------| longest\n\necho ${stringZ%b*c} # abcABC123ABCa\n# Strip out shortest match between 'b' and 'c', from back of $stringZ.\n\necho ${stringZ%%b*c} # a\n# Strip out longest match between 'b' and 'c', from back of $stringZ.<\/pre>\n<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\nThis operator is useful for generating filenames.<\/p>\n \n Example 10-3. Converting graphic file formats, with filename change<\/b><\/p>\n\n\n\n\n#!\/bin\/bash\n# cvt.sh:\n# Converts all the MacPaint image files in a directory to \"pbm\" format.\n\n# Uses the \"macptopbm\" binary from the \"netpbm\" package,\n#+ which is maintained by Brian Henderson (bryanh@giraffe-data.com).\n# Netpbm is a standard part of most Linux distros.\n\nOPERATION=macptopbm\nSUFFIX=pbm # New filename suffix. \n\nif [ -n \"$1\" ]\nthen\n directory=$1 # If directory name given as a script argument...\nelse\n directory=$PWD # Otherwise use current working directory.\nfi \n\n# Assumes all files in the target directory are MacPaint image files,\n#+ with a \".mac\" filename suffix.\n\nfor file in $directory\/* # Filename globbing.\ndo\n filename=${file%.*c} # Strip \".mac\" suffix off filename\n #+ ('.*c' matches everything\n\t\t\t #+ between '.' and 'c', inclusive).\n $OPERATION $file > \"$filename.$SUFFIX\"\n # Redirect conversion to new filename.\n rm -f $file # Delete original files after converting. \n echo \"$filename.$SUFFIX\" # Log what is happening to stdout.\ndone\n\nexit 0\n\n# Exercise:\n# --------\n# As it stands, this script converts *all* the files in the current\n#+ working directory.\n# Modify it to work *only* on files with a \".mac\" suffix.<\/pre>\n<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/div>\n\n Example 10-4. Converting streaming audio files to\u00a0ogg<\/i><\/b><\/p>\n\n\n\n\n#!\/bin\/bash\n# ra2ogg.sh: Convert streaming audio files (*.ra) to ogg.\n\n# Uses the \"mplayer\" media player program:\n# http:\/\/www.mplayerhq.hu\/homepage\n# Uses the \"ogg\" library and \"oggenc\":\n# http:\/\/www.xiph.org\/\n#\n# This script may need appropriate codecs installed, such as sipr.so ...\n# Possibly also the compat-libstdc++ package.\n\nOFILEPREF=${1%%ra} # Strip off the \"ra\" suffix.\nOFILESUFF=wav # Suffix for wav file.\nOUTFILE=\"$OFILEPREF\"\"$OFILESUFF\"\nE_NOARGS=85\n\nif [ -z \"$1\" ] # Must specify a filename to convert.\nthen\n echo \"Usage: `basename $0` [filename]\"\n exit $E_NOARGS\nfi\n\n##########################################################################\nmplayer \"$1\" -ao pcm:file=$OUTFILE\noggenc \"$OUTFILE\" # Correct file extension automatically added by oggenc.\n##########################################################################\n\nrm \"$OUTFILE\" # Delete intermediate *.wav file.\n # If you want to keep it, comment out above line.\n\nexit $?\n\n# Note:\n# ----\n# On a Website, simply clicking on a *.ram streaming audio file\n#+ usually only downloads the URL of the actual *.ra audio file.\n# You can then use \"wget\" or something similar\n#+ to download the *.ra file itself.\n\n# Exercises:\n# ---------\n# As is, this script converts only *.ra filenames.\n# Add flexibility by permitting use of *.ram and other filenames.\n#\n# If you're really ambitious, expand the script\n#+ to do automatic downloads and conversions of streaming audio files.\n# Given a URL, batch download streaming audio files (using \"wget\")\n#+ and convert them on the fly.<\/pre>\n<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/div>\nA simple emulation of\u00a0getopt\u00a0using substring-extraction constructs.<\/p>\n \n Example 10-5. Emulating\u00a0getopt<\/i><\/b><\/p>\n\n\n\n\n#!\/bin\/bash\n# getopt-simple.sh\n# Author: Chris Morgan\n# Used in the ABS Guide with permission.\n\ngetopt_simple()\n{\n echo \"getopt_simple()\"\n echo \"Parameters are '$*'\"\n until [ -z \"$1\" ]\n do\n echo \"Processing parameter of: '$1'\"\n if [ ${1:0:1} = '\/' ]\n then\n tmp=${1:1} # Strip off leading '\/' . . .\n parameter=${tmp%%=*} # Extract name.\n value=${tmp##*=} # Extract value.\n echo \"Parameter: '$parameter', value: '$value'\"\n eval $parameter=$value\n fi\n shift\n done\n}\n\n# Pass all options to getopt_simple().\ngetopt_simple $*\n\necho \"test is '$test'\"\necho \"test2 is '$test2'\"\n\nexit 0 # See also, UseGetOpt.sh, a modified version of this script.\n\n---\n\nsh getopt_example.sh \/test=value1 \/test2=value2\n\nParameters are '\/test=value1 \/test2=value2'\nProcessing parameter of: '\/test=value1'\nParameter: 'test', value: 'value1'\nProcessing parameter of: '\/test2=value2'\nParameter: 'test2', value: 'value2'\ntest is 'value1'\ntest2 is 'value2'<\/pre>\n<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/div>\n<\/dd>\n<\/dl>\n<\/div>\n\n Substring Replacement<\/b><\/p>\n\n- ${string\/substring\/replacement}<\/dt>\n
- Replace first\u00a0match<\/i>\u00a0of\u00a0$substring<\/i><\/tt>\u00a0with\u00a0$replacement<\/i><\/tt>.\u00a0[2]<\/dd>\n
- ${string\/\/substring\/replacement}<\/dt>\n
- Replace all matches of\u00a0$substring<\/i><\/tt>\u00a0with\u00a0$replacement<\/i><\/tt>.<\/p>\n
\n\n\n\nstringZ=abcABC123ABCabc\n\necho ${stringZ\/abc\/xyz} # xyzABC123ABCabc\n # Replaces first match of 'abc' with 'xyz'.\n\necho ${stringZ\/\/abc\/xyz} # xyzABC123ABCxyz\n # Replaces all matches of 'abc' with # 'xyz'.\n\necho ---------------\necho \"$stringZ\" # abcABC123ABCabc\necho ---------------\n # The string itself is not altered!\n\n# Can the match and replacement strings be parameterized?\nmatch=abc\nrepl=000\necho ${stringZ\/$match\/$repl} # 000ABC123ABCabc\n# ^ ^ ^^^\necho ${stringZ\/\/$match\/$repl} # 000ABC123ABC000\n# Yes! ^ ^ ^^^ ^^^\n\necho\n\n# What happens if no $replacement string is supplied?\necho ${stringZ\/abc} # ABC123ABCabc\necho ${stringZ\/\/abc} # ABC123ABC\n# A simple deletion takes place.<\/pre>\n<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/dd>\n- ${string\/#substring\/replacement}<\/dt>\n
- If\u00a0$substring<\/i><\/tt>\u00a0matches\u00a0front<\/em>\u00a0end of\u00a0$string<\/i><\/tt>, substitute\u00a0$replacement<\/i><\/tt>\u00a0for\u00a0$substring<\/i><\/tt>.<\/dd>\n
- ${string\/%substring\/replacement}<\/dt>\n
- If\u00a0$substring<\/i><\/tt>\u00a0matches\u00a0back<\/em>\u00a0end of\u00a0$string<\/i><\/tt>, substitute\u00a0$replacement<\/i><\/tt>\u00a0for\u00a0$substring<\/i><\/tt>.<\/p>\n
\n\n\n\nstringZ=abcABC123ABCabc\n\necho ${stringZ\/#abc\/XYZ} # XYZABC123ABCabc\n # Replaces front-end match of 'abc' with 'XYZ'.\n\necho ${stringZ\/%abc\/XYZ} # abcABC123ABCXYZ\n # Replaces back-end match of 'abc' with 'XYZ'.<\/pre>\n<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n\n Go to- | | | | | | | | | | | | | | | | | |