Lecture Nine

Regular Expressions

Regular Expressions

• many Unix utilities use regular expressions: grep, sed, awk, vi, perl, Tcl
• shell filename matches are not regular expressions (eg. *.c)
• examples in this lecture will use the grep utility and the file cars
• regular expressions are used to search for or match text:
  • literal text can be used to search for that text
    • grep "chevy" cars
  • . matches any character (similar to ? wildcard)
    • grep ".c" cars
    • grep "5..." cars
  • [ ] matches any character within the square brackets (similar to [ ] wildcard)
    • grep "[cC]hevy" cars
    • grep "[0-9][0-9][0-9][0-9][0-9]" cars
  • matches any character not within the square brackets (similar to [! ] wildcard)
    • grep "ford" cars
  • ^ matches beginning of line
    • grep "^f" cars
  • $ matches end of line
    • grep " [0-9][0-9][0-9]$" cars
  • * following any character denotes zero or more occurrences of that character
    • grep "ford.*83" cars
    • grep "^ 65" cars
  • \ inhibits meaning of special characters
    • grep ' [0-9][0-9][0-9]\/tt> cars
• extended regular expressions are not recognized by grep, can use egrep or grep -E:
  • ( reg-exp ) parentheses used for grouping
    • egrep "^( + +){2}65" cars
    • egrep "ford" cars
  • | means OR, matches reg-exp on either side of the vertical bar
    • egrep "ford|chevy" cars
• regular expression characters may or may not need to be escaped - varies from program to program
  • eg. egrep and awk use ( and ) for grouping, sed uses ( and ) unless -r option is used
• regular expressions may or may not need delimiters - varies from program to program
  • eg. grep and egrep don't use delimiters,sed and awk use delimiters, eg. /string/
• other examples of regular expressions
  (Mr|Mrs) Smith - match either "Mr Smith" or "Mrs Smith"
  [a-zA-Z]+ - match one or more letters
  ^[a-zA-Z]*$ - match lines with only letters
  0-9+ - match string not containing digits
  [+-]?([0-9]+[.]?[0-9]*|[.][0-9]+)([eE][+-]?[0-9]+)? - match valid "C" programming numbers

grep

• uses regular expression for pattern, eg. grep 'reg-exp' filename, then prints matched lines
• gives 0 exit status if pattern matched
• options:
  • -c - counts matched lines instead of printing them
  • -i - ignores case
  • -n - precedes each line with a line number
  • -v - reverses sense of test, eg. finds lines not matching pattern
• examples, using the file cars
  • grep 'chevy' cars - display only lines containing the string "chevy"
  • grep -c 'chevy' cars - display count of lines containing the string "chevy"
  • grep -i 'chevy' cars - display only lines containing the string "chevy", ignoring case
  • grep -ic 'chevy' cars - display count of lines containing the string "chevy", ignoring case
  • grep -v 'chevy' cars - display only lines not containing the string "chevy"
  • grep -ivc 'chevy' cars - display count of lines not containing the string "chevy", ignoring case
  • grep -n 'chevy' cars - display only lines containing the string "chevy", with line numbers