in reply to Pulling out pairs

while (<>) {
    chomp;
    (my ($key, $val) = split /:/) == 2 or next;
}

[download]

    -- Chip Salzenberg, Free-Floating Agent of Chaos

Replies are listed 'Best First'.
Re: Re: Pulling out pairs
by blakem (Monsignor) on Dec 13, 2001 at 07:44 UTC
    Thats rather clever.... There is a subtle issue with extrapolating that particular idiom though. Suppose we wanted to match lines of exactly four fields, but were only interested in the first two fields of such lines: 
    dog:cat:pig       <== skip, only has three fields
Amy:Ann           <== skip, only has two fields
ape:ant:bug:car   <== match this line (it has four fields)
                      but we only need to keep 'ape' and 'ant'

    [download]

    Tweaking the code above, we might think that this will do the trick for us: 

    while (<DATA>) {
    chomp;
    (my ($first, $second) = split /:/) == 4 or next;
    print "$first $second\n";
}

__DATA__
dog:cat:pig
Amy:Ann
ape:ant:bug:car

    [download]
    However, that doesn't seem to work for us, nothing gets printed. In fact, try as we might, we can't construct any line at all that passes the IDIOM==4 test.

    So, does anyone want to guess why IDIOM==4 is so different from IDIOM==2?

    Update: As an added hint, testing for IDIOM==3 will match the 'dog:cat:pig' line....

    -Blake

[reply]
[d/l]
[select]
      The problem is that the optimiser will see: 
      ($a, $b) = split /:/

      [download]
      And know that things can go faster by turning that into: 
      ($a, $b) = split /:/, $_, 3

      [download]
      thus you actually should write: 
      (my ($a, $b) = split /:/, $_, 5) == 4 or next;

      [download]
[reply]
[d/l]
[select]
        Well done. That's exactly what I was getting at.

        (I would have just /msged this response, except you cant /msg anonymonk....)

        -Blake

[reply]

In Section Seekers of Perl Wisdom