如何删除新行字符，直到每行都有特定数量的特定字符实例？

Question

我有一个非常混乱的管道分隔文件，我需要将其加载到数据库中。该文件有 35 个字段，因此有 34 个管道。其中一个字段由 HTML 代码组成，对于某些记录，该代码包含多个换行符。不幸的是，没有关于断线在哪里的说法。

我提出的解决方案是计算每行中管道的数量，直到该数字达到 34，从该行中删除新行字符。我对 Perl 并不是非常精通，但我认为我已经接近实现我想要做的事情了。有什么建议吗？

#!/usr/local/bin/perl

use strict;

open (FILE, 'test.txt');

while (<FILE>) {
    chomp;
    my $line = $_;
    #remove null characters that are included in file
    $line =~ tr/\x00//;
    #count number of pipes
    my $count = ($line =~ tr/|//);
    #each line should have 34 pipes
    while ($count < 34) {
        #remove new lines until line has 34 pipes
        $line =~ tr/\r\n//;
        $count = ($line =~ tr/|//);
        print "$line\n";
    }
}

Answer 1

我想这应该有效。

#!/usr/bin/perl

use strict;

open (FILE, 'test.txt');

my $num_pipes = 0, my $line_num = 0;
my $tmp = "";
while (<FILE>)
{
    $line_num++;
    chomp;
    my $line = $_;
    $line =~ tr/\x00//; #remove null characters that are included in file
    $num_pipes += ($line =~ tr/|//); #count number of pipes
    if ($num_pipes == 34 && length($tmp))
    {
            $tmp .= $line;
            print "$tmp\n";
            # Reset values.
            $tmp = "";
            $num_pipes = 0;
    }
    elsif ($num_pipes == 34 && length($tmp) == 0)
    {
            print "$line\n";
            $num_pipes = 0;
    }
    elsif ($num_pipes < 34)
    {
            $tmp .= $line;
    }
    elsif ($num_pipes > 34)
    {
            print STDERR "Error before line $line_num. Too many pipes ($num_pipes)\n";
            $num_pipes = 0;
            $tmp = "";
    }
}

Answer 2

摆弄 $/，输入记录分隔符？

while (!eof(FILE)) {

    # assemble a row of data: 35 pipe separated fields, possibly over many lines
    my @fields = ();
    {
        # read 34 fields from FILE:
        local $/ = '|';
        for (1..34) {
            push @fields, scalar <FILE>;
        }
    }   # $/ is set back to original value ("\n") at the end of this block

    push @fields, scalar <FILE>;  # read last field, which ends with newline
    my $line = join '|', @fields;
    ... now you can process $line, and you already have the @fields ......
}