G'day spstansbury,

I see Tux has provided you with a better way to tackle your task. The reason you're getting the results you see with your posted solution is as follows.

Given this test.csv (newlines shown as "$" and tabs as "^I"):

$ cat -vet test.csv
quote at end"$
no quote at end$
quote and space at end" $
quote and tab at end"^I$
[download]

Your posted code matches every line:

$ perl -ne 'print $_ if $_ =~ /[^"]$/' < test.csv
quote at end"
no quote at end
quote and space at end" 
quote and tab at end"
[download]

The reason for this becomes apparent when you look at exactly what $_ contains:

$ perl -ne 'print ">>>$_<<<\n" if $_ =~ /[^"]$/' < test.csv
>>>quote at end"
<<<
>>>no quote at end
<<<
>>>quote and space at end" 
<<<
>>>quote and tab at end"    
<<<
[download]

Each line ends with a newline which, not being a double-quote, means /[^"]$/ matches.

You can get around this most easily but simply using the -l switch:

$ perl -lne 'print $_ if $_ =~ /[^"]$/' < test.csv
no quote at end
quote and space at end" 
quote and tab at end"
[download]

If you look at perlrun, you'll see that's equivalent to chomping each input line and appending a newline (or whatever the current value of $/ is) to each output line:

$ perl -ne 'chomp; print "$_\n" if $_ =~ /[^"]$/' < test.csv
no quote at end
quote and space at end" 
quote and tab at end"
[download]

[Note: the -l switch is a little more complicated than my simplistic explanation - see the doco for details.]