Globbing

• Linux shell supports globbing patterns as a means of matching filenames - globbing patterns support several wildcard characters/syntax to create patterns that are then checked against filenames
  • ? - match any one character
    • e.g. ?.txt - match file name with any one character that is then followed by .txt
    • to match two any two characters, you can use two ? symbo;s
    • e.g. CS???.txt - matches any filenames beginning with CS followed by any 3 characters followed by .txt
  • • • match 0+ characters
      • e.g. *.txt - match any file name ending with .txt
  • [] - specify one character from a set of characters
    • e.g. [aeiou].txt - matches any filename that has a filename of a.txt, e.txt, etc, followed by .txt
    • character ranges can also be specified, e.g. [1-9], [a-z]
  • [!] - match any one character aside from those specified in set
    • e.g. [!aeiou].txt - match any single character file name that is not vowel
  • {} - technically not a globbing pattern, but means of saying “or”
    • e.g. *.{cc, cpp} - match all .cc, .ccp files
• To suppress globbing capability, use single/double quotes around string
• one application is using find - can use globbing pattern to search for files
  • use quotes for example, in find /usr/share/dict -name "w*" - direct shell to not use globbing, forward globbing pattern to find itself

Pipes

• Linux shell supports ability to redirect output generated by one program to the output of another program - achieved with a pipe
  • Linux pipe used to connect standard output stream of a program to the standard input stream of a program that executes right after the termination of the first program
• e.g. get first 20 lines of a file, find word count
  • one solution: run head -n 20 sample.txt > first20.txt and wc -w < first20.txt
  • using pipe: head -n 20 sample.txt | wc -w, or cat sample.txt | head -n 20 | wc -w - results from sample.txt are piped into standard input stream of head
• e.g. Print duplicate-free list of words occuring in any files words*.txt
  • cat words*.txt | sort | uniq

Embedded Commands

• We can also embed the output of a command in a string - provide output of command as an argument to another program/command
• e.g. echo "It is currently $(date) and I am $(whoami)"
• When double quotes are used, a single command-line argument is provided to echo, otherwise are are supplying multiple whitespace-separated arguments
• To suppress the execution of embedded commands, we can place text in single quotes (e.g. echo 'It is currently $(date)' , or we could also do backticks surrounding date (date)

egrep

• It is often to search for the contents within text files, so Linux shell provides egrep utility
• Tool searches given files and outputs lines that match 1+ patterns
• Patterns provided as (extended) regular expressions - not the same as globbing patterns
• e.g. egrep CS136 sample.txt - run egrep with regular expression CS136 as first argument and filename sample.txt as second argument
  • could also write command as ca sample.txt | egrep CS136 - egrep accepts text as input on standard input stream along with command line argument filenames
• egrep will independently try to match the regular expression on each line of the provided file/text
  • Line - sequence of 1+ characters where last character in sequence is \n