http://www.perlmonks.org?node_id=435223


in reply to Re: Re: Perl Idioms Explained - keys %{{map{$_=>1}@list}}
in thread Perl Idioms Explained - keys %{{map{$_=>1}@list}}

undef is not documented to work on slices. That command should probably generate at least a warning.

Caution: Contents may have been coded under pressure.

Replies are listed 'Best First'.
Re^4: Perl Idioms Explained - keys %{{map{$_=>1}@list}}
by pilcrow (Sexton) on Mar 02, 2005 at 06:54 UTC
    perlfunc's entry for undef suggests this behavior on slices, if one squints:
    Undefines the value of EXPR, which must be an lvalue... Saying undef $hash{$key} will probably not do what you expect...
    That is, taking "undef EXPR" as (usually) equivalent to "EXPR = undef", the behavior is sensible, even expected. At any rate, a warning seems inappropriate, as this feature is long-standing:
    $ perl -le 'undef @h{qw|a b c|}; print "$]: ", join " ", keys %h' 5.00404: a b c
      That is, taking "undef EXPR" as (usually) equivalent to "EXPR = undef", the behavior is sensible, even expected.
      But that isn't what it does.
      my @ar = (1..10); undef @ar[0..5]; print "@ar\n"; # 1 2 3 4 5 7 8 9 10 @ar[0..5] = undef; print "@ar\n"; # 7 8 9 10
      The former interprets the slice as a scalar list, and undefs the last element only. The reason all the keys get defined is that it puts the slice in lvalue context, and autovivification occurs. The latter is equivalent to assigning a list to a list.

      Caution: Contents may have been coded under pressure.
        The former interprets the slice as a scalar list, and undefs the last element only. The reason all the keys get defined is that it puts the slice in lvalue context, and autovivification occurs. The latter is equivalent to assigning a list to a list.
        I think I see your point. Would you then have expected, in nature, the following to be equivalent:
        $h{c} = undef; # 1 undef $h{c}; # 2 undef @h{qw|a b c|}; # 3
        with no. 3 taking the hash slice as a plain list, and that in a scalar context?