http://www.perlmonks.org?node_id=883994

MeatLips has asked for the wisdom of the Perl Monks concerning the following question:

Here's a question for the wise monks... I have this single line string:

key1=val1 key2=val2 key3=val3 key4="val4a val4b" key5="val5key=(0 1 2 3)" key6=(val6a val6b)

I want to parse that into an array so I can have something simple like this:

foreach my $x (@array) {
    print "$x\n";
}

[download]

return the following: 

key1=val1
key2=val2
key3=val3
key4="val4a val4b"
key5="valkey=(0 1 2 3)"
key6=(val6a val6b)

[download]

I've been tearing my hair out, poring over various regex texts, trying various ways to use 'split', anything to figure out how to split this pig up the way I want it. What would the wise monks here suggest?

Replies are listed 'Best First'.
Re: string parsing with split
by BrowserUk (Patriarch) on Jan 24, 2011 at 20:42 UTC

    Use a lookahead:

    $s = q[key1=val1 key2=val2 key3=val3 key4="val4a val4b" key5="val5key=
+(0 1 2 3)" key6=(val6a val6b)];;

@a = split " (?=key)", $s;;

print for @a;;
key1=val1
key2=val2
key3=val3
key4="val4a val4b"
key5="val5key=(0 1 2 3)"
key6=(val6a val6b)

    [download]
    Examine what is said, not who speaks -- Silence betokens consent -- Love the truth but pardon error.
    "Science is about questioning the status quo. Questioning authority".
    In the absence of evidence, opinion is indistinguishable from prejudice.
[reply]
[d/l]

      This would seem to work if I literally had the word "key" in the string. I was using that as an example. The actual string has different words that are not the same.

      Here's something that looks a bit closer to the actual string:

      platform=linux hpfFamily=hwseries npsFamily=SWseries rackcnt=1 SPAs=1 SPUpSPA=6 CPUs=6 shrParts="/part1 /home/part2" maintIface="eth2" DEncpSPA=4 nDEncs=4 npsMName="SystemName" portnums="ports=(0 1 2 3)" ints=(eth0 eth1)
[reply]
[d/l]
        Here's something that looks a bit closer to the actual string:

        Then here's something that may be a bit closer to a working solution for you? 

        $s = q[platform=linux hpfFamily=hwseries npsFamily=SWseries rackcnt=1 
+SPAs=1 SPUpSPA=6 CPUs=6 shrParts="/part1 /home/part2" maintIface="eth
+2" DEncpSPA=4 nDEncs=4 npsMName="SystemName" portnums="ports=(0 1 2 3
+)" ints=(eth0 eth1)];;

print for split ' (?=\w+=)', $s;;
platform=linux
hpfFamily=hwseries
npsFamily=SWseries
rackcnt=1
SPAs=1
SPUpSPA=6
CPUs=6
shrParts="/part1 /home/part2"
maintIface="eth2"
DEncpSPA=4
nDEncs=4
npsMName="SystemName"
portnums="ports=(0 1 2 3)"
ints=(eth0 eth1)

        [download]
        Examine what is said, not who speaks -- Silence betokens consent -- Love the truth but pardon error.
        "Science is about questioning the status quo. Questioning authority".
        In the absence of evidence, opinion is indistinguishable from prejudice.
[reply]
[d/l]
Re: string parsing with split
by ikegami (Patriarch) on Jan 24, 2011 at 20:58 UTC
    Another way, 
    my @pairs = /
   \G \s* (
      [^=]+
      =
      (?: "  [^"]* "
      |   \( [^)]* \)
      |   \S*
      )
   )
/xg;

    [download]
[reply]
[d/l]
Re: string parsing with split
by AnomalousMonk (Archbishop) on Jan 24, 2011 at 22:27 UTC

    Here's my take on the problem. It has the advantage, IMHO, of being more easily adaptable to changing requirements because it is more modular.

    Notes:

    • The regex uses  \x22 in place of  " (double-quote) to avoid Windoze command-line escape-ology.
    • A quoted string cannot contain any sort of double-quote, escaped or otherwise.
    • A parenthetic group cannot contain a  ')' (right-paren).
    (Sorry for the line-wrap.) 

    >perl -wMstrict -le
"my $s = 'key1=val1 key2=val2 key3=val3 key4=\"val4a val4b\" '
       . 'key5=\"val5key=(0 1 2 3)\" key6=(val6a val6b)'
       ;
 ;;
 my $key   = qr{ [[:alpha:]] [[:alnum:]]+ }xms;
 my $val   = qr{ [[:alpha:]] [[:alnum:]]+ }xms;
 my $d_quo = qr{ \x22 [^\x22]* \x22 }xms;
 my $paren = qr{ [(]  [^)]*    [)]  }xms;
 ;;
 my $vals  = qr{ $val | $d_quo | $paren }xms;
 ;;
 my @opts = $s =~ m{ $key \s* = \s* $vals }xmsg;
 ;;
 print qq{'$s'};
 print qq{'$_'} for @opts;
"
'key1=val1 key2=val2 key3=val3 key4="val4a val4b" key5="val5key=(0 1 2
+ 3)" key6=(val6a val6b)'
'key1=val1'
'key2=val2'
'key3=val3'
'key4="val4a val4b"'
'key5="val5key=(0 1 2 3)"'
'key6=(val6a val6b)'

    [download]
[reply]
[d/l]
[select]
Re: string parsing with split
by jethro (Monsignor) on Jan 24, 2011 at 20:57 UTC
    I think Text::CSV is a well known solution to such problems
[reply]
      It's not up to the task. It can be configured to allow quoting to start in the middle of a field (loose quotes), but it doesn't support multiple quoting characters or balanced quotes (start quote character (e.g. "(") different than end quote character (")")).
[reply]

Back to Seekers of Perl Wisdom