Perl Cheatsheet 3

August 30, 2014

RegEx cont.

Variables

  • Match vars.
    • if regex have parens i.e. groups
    • if group got successful match
      • then the content of matched pattern stored in $1, meaning first match against first group
      • e.g. /(\S+)(\S+),(\S+)/ corresponding matched content is print $1 $2 $3
    • classical match idiom: if(xx =~ xx) do sth else die
    • three special vars.
      • a regex matching a string divides the string into 3 pars
      • $&: the part actually matched
      • $\backslash: the part before matched
      • $': the part after matched
  • Noncapturing parens
    • sometimes, we use parens only for precedence, however, it still trigger matching variable.
    • how can we escape from that, to tell perl not to trigger matching var in parens ?
    • to avoid triggering memory, change (…) to (?:…)
    • e.g. /(\S+)(?:and|or)(\S+)/ the middle parens won’t be counted as $2, instead, the last parens is $2.
  • Named captures
    • even better way, to tell which parens I want to capture, and give name to that matching var.
    • e.g. /(?<name1>\S+)(lorem)/ now print "$+{name1}" will print the content of matching \S+ in the named capture you create.

Substitution

  • s/old/new/
    • old is RegEx
    • new is treated as double quoted string, so if there is $var inside, will be replaced as its value. never think of putting RegEx in new!
    • RegEx triggers back reference var: s/with(\w+)/without$1/ like said before, $1 is special var in double quoted string context, will be replaced by its value
    • idiom: if (s/old/new/) {…}
  • Alternate delimiter
    • e.g. #^https://#http://# again, don’t put ^in http, it’s merely treated as “string”!
  • More Modes
    • Global matching, s/\s+/ /g transform to canonical whitespace
    • case-insensitive, s/old/new/i
    • . including new lines, s/old/new/s
  • List-context match
    • you can store parens matched result in a list
    • e.g. my($first, $second, $third) = /(\S)(\S),(\S)/;
    • e.g. my @words = /(\w+)/g
  • **Non-greedy quantifier
    • quantifier normally go long, then back off
    • we can tell perl go as short as you can !
    • Append ? to quantifier to make it lazy
  • Split function
    • my @fields = split /separator/, $string
    • separator here is RegEx
    • my @ip = split /\./, "192.168.1.1";
  • Case shifting
    • \U: take every var following \U, make them uppercase
    • \L: take every var following \L, make them lowercase
    • \u: take first char of followed, make it uppercase
    • \l: take first char of followed, make it lowercase
    • \E: where case shifting stops
    • N.B. all these happen in “string”!
    • e.g. s/(\w+)/\U$1/
    • e.g. s/(\w{2})/\u\L$1/ig initial cap