Surprise: scalar(($x, $y) = split)

I was trying to read a text file each of whose lines should contain two whitespace-separated fields. I wrote the file so I know it has two fields per line, but it's better to be cautious, so I decided to check that it's indeed two fields. I wrote this.

while (<>) {
    2 == (my($x, $y) = split " ") or die "wrong number of fields";
    ...
}
[download]

That gives the error

wrong number of fields at -e line 1, <> line 1.
[download]

despite that line 1 of the file really has exactly two fields. Why? Solution under the fold.

Let's look more closely.

$ perl -e 'warn 0+(@x = split " ", "foo bar\n");'
2 at -e line 1.
$ perl -e 'warn 0+(($x, $y) = split " ", "foo bar\n");'
3 at -e line 1.
$
[download]

The reason is one of the several ways how the split function is magical. Let me quote the relevant phrase from the perlfunc section about split.

When assigning to a list, if LIMIT is omitted, or zero, Perl supplies a LIMIT one larger than the number of variables in the list, to avoid unnecessary work.

The side effect here could be a feature or a misfeature, but in any case, this is what perl 5.10, 5.12 and 5.14 does.

Comment on Surprise: scalar(($x, $y) = split) Select or Download Code

Replies are listed 'Best First'.
Re: Surprise: scalar(($x, $y) = split) by JavaFan (Canon) on Oct 28, 2011 at 14:55 UTC
`$ perl -e 'warn 0+(@x = split " ", "foo bar\n");' 2 at -e line 1. $ perl -e 'warn 0+(($x, $y) = split " ", "foo bar\n");' 3 at -e line 1. $` [download] Note that there are three magical features happening. First there's the magic of the first argument; if the first argument of split is a pattern consisting of a single space, split splits on any whitespace. Which includes the newline. So, you actually have 3 fields, ("foo", "bar", "" -- the empty string following the newline), so it's the first case that's odd, not the second. It's the second magical feature that's to blame - in the absence of a limit, trailing empty fields are removed. You wouldn't have been surprised if you had chomped your line.	[reply] [d/l]
Re^2: Surprise: scalar(($x, $y) = split) by ambrus (Abbot) on Oct 28, 2011 at 15:24 UTC
The other two magic might be confusing but they're not really necessary here. I can still write this and be surprised. `$ perl -we '$n = () = split ",", "one,two,three,four,five"; print "I f +ound $n comma-separated words\n"' I found 1 comma-separated words $ perl -we '$n = ($x) = split ",", "one,two,three,four,five"; print "I + found $n comma-separated words, the first of which is $x\n"' I found 2 comma-separated words, the first of which is one` [download] And even if I chomp the line, there could be trailing spaces or carriage returns in the line that I expect `split " "` to cut off.	[reply] [d/l] [select]
Re^2: Surprise: scalar(($x, $y) = split) by hbm (Hermit) on Oct 28, 2011 at 15:26 UTC
Adding to the confusion? `$ perl -e 'warn 0+(($x, $y) = split/ /, "foo bar\n");' 2 at -e line 1. $ perl -e 'warn 0+(($x, $y) = split$", "foo bar\n");' 2 at -e line 1.` [download]	[reply] [d/l]
Re: Surprise: scalar(($x, $y) = split) by roboticus (Chancellor) on Oct 28, 2011 at 14:03 UTC
Update: Everything below is wrong, as ambrus mentions in the reply. (Note to self: Self, just because you get the correct result doesn't mean that you got the right answer!) ambrus: You forgot some parenthesis: `(2==(my($x,$y) = split " ")) or die "wrong number of fieldsd";` Example: $ cat foo.pl use strict; use warnings; my $x; my $y; while (<DATA>) { chomp; (2==0+(($x, $y) = split /:/,$_)) or print "wrong number of fields\ +n"; print "<$_>, <$x>, <$y>\n"; } __DATA__ FOO:BAR:BAZ:PLOVE FOO:BAR:BAZ FOO:BAR BAZ $ perl foo.pl wrong number of fields <FOO:BAR:BAZ:PLOVE>, <FOO>, <BAR> wrong number of fields <FOO:BAR:BAZ>, <FOO>, <BAR> <FOO:BAR>, <FOO>, <BAR> wrong number of fields Use of uninitialized value $y in concatenation (.) or string at foo.pl + line 7, <DATA> line 4. <BAZ>, <BAZ>, <> [download] ...roboticus When your only tool is a hammer, all problems look like your thumb.	[reply] [d/l] [select]
Re^2: Surprise: scalar(($x, $y) = split) by ambrus (Abbot) on Oct 28, 2011 at 14:05 UTC
It's not the parenthesis that helps there but the chomp. `$ perl -e 'warn 0+(($x, $y) = split " ", "foo bar\n");' 3 at -e line 1. $ perl -e 'warn 0+(($x, $y) = split " ", "foo bar");' 2 at -e line 1.` [download]	[reply] [d/l]
Re^3: Surprise: scalar(($x, $y) = split) by roboticus (Chancellor) on Oct 28, 2011 at 14:21 UTC
ambrus: Ha, ha--color me surprised as well! Corrected example: $ cat foo.pl use strict; use warnings; my $x; my $y; while (<DATA>) { chomp; 2==(($x, $y) = split " ") or print "wrong number of fields\n"; print "<$_>, <$x>, <$y>\n"; } __DATA__ FOO BAR BAZ PLOVE FOO BAR BAZ FOO BAR BAZ $ perl foo.pl wrong number of fields <FOO BAR BAZ PLOVE>, <FOO>, <BAR> wrong number of fields <FOO BAR BAZ>, <FOO>, <BAR> <FOO BAR>, <FOO>, <BAR> wrong number of fields Use of uninitialized value $y in concatenation (.) or string at foo.pl + line 7, <DATA> line 4. <BAZ>, <BAZ>, <> [download] ...roboticus When your only tool is a hammer, all problems look like your thumb.	[reply] [d/l]
Re: Surprise: scalar(($x, $y) = split) by Fox (Pilgrim) on Oct 28, 2011 at 18:17 UTC
Not so pretty.. but anyway: `$ perl -wle '(my($x,$y) = @{[]} = split/ /,"foo bar") == 2 or die "boo +";warn "[$x] [$y]"' [foo] [bar] at -e line 1. $ perl -wle '(my($x,$y) = @{[]} = split/ /,"foo bar baz") == 2 or die +"boo";warn "[$x] [$y]"' boo at -e line 1.` [download]	[reply] [d/l]
Re: Surprise: scalar(($x, $y) = split) by Anonymous Monk on Oct 29, 2011 at 08:19 UTC
use strict; use warnings; { my( $x , $y ) ; print 'scalar( ($x, $y ) ) ', scalar( ($x, $y ) ) +, "\n"; print 'scalar( ($x, $y ) ) = () ', scalar( ($x, $y ) = + () ), "\n"; print 'scalar( ($x, $y ) ) = (undef) ', scalar( ($x, $y ) = + (undef) ), "\n"; print 'scalar( ($x, $y ) ) = (undef,undef) ', scalar( ($x, $y ) = + (undef,undef) ), "\n"; print 'scalar( ($x, $y ) ) = (0) ', scalar( ($x, $y ) = + (0) ), "\n"; print 'scalar( ($x, $y ) ) = (0,0) ', scalar( ($x, $y ) = + (0,0) ), "\n"; print 'scalar( ($x, $y ) ) = (1,0) ', scalar( ($x, $y ) = + (1,0) ), "\n"; print 'scalar( ($x, $y ) ) = (1,1) ', scalar( ($x, $y ) = + (1,1) ), "\n"; print 'scalar( ($x, $y ) ) = (1,1,1) ', scalar( ($x, $y ) = + (1,1,1) ), "\n"; print "\n\n"; } __END__ Useless use of private variable in void context at - line 5. Use of uninitialized value $y in print at - line 5. scalar( ($x, $y ) ) scalar( ($x, $y ) ) = () 0 scalar( ($x, $y ) ) = (undef) 1 scalar( ($x, $y ) ) = (undef,undef) 2 scalar( ($x, $y ) ) = (0) 1 scalar( ($x, $y ) ) = (0,0) 2 scalar( ($x, $y ) ) = (1,0) 2 scalar( ($x, $y ) ) = (1,1) 2 scalar( ($x, $y ) ) = (1,1,1) 3 [download]	[reply] [d/l]