Perl正则表达式存储数组中的匹配项

Question

I have a file with strings in each row as follows

"229269_2,190594_2,94552_2,266076_2,269628_2,165328_2,99319_2,263339_2,263300_2,99315_2,271509_2,2714",A,1 the next line could look like 84545,X,2

I'm trying to parse this text in Perl. Note: quotes are present in the strings when there are several of them in a row, but not present if there is only item I would like to parse each item into an array. I tried the following regex

@fields = ($_ =~  /(\d+\_\d+),*/g);

but it is missing the last 2714 . How do I capture that edge case? Any help appreciated. Thanks in advance

Answer 1

It looks like you have a CSV File, so use an actual CSV parser for it like Text::CSV .

After you parse the columns, you can separate your first field into the array:

use strict;
use warnings;

use Text::CSV;

my $csv = Text::CSV->new ( { binary => 1 } )  # should set binary attribute.
    or die "Cannot use CSV: ".Text::CSV->error_diag ();

my $line = qq{"229269_2,190594_2,94552_2,266076_2,269628_2,165328_2,99319_2,263339_2,263300_2,99315_2,271509_2,2714",A,1 the next line could look like 84545,X,2};

if ($csv->parse($line)) {
    my @columns = $csv->fields();
    my @nums = split ',', $columns[0];

    print "@nums\n";
}

Outputs:

229269_2 190594_2 94552_2 266076_2 269628_2 165328_2 99319_2 263339_2 263300_2 99315_2 271509_2 2714

Why not a regex ?

Yes, of course it's possible to use a regex for practically anything. But what you need to understand is that this will make your code extremely fragile and difficult to maintain.

Even if you want to use a regular expression, you should STILL do this in two steps. First separate the initial column(s) of your CSV, and then process the specific column that you're worried about.

Because you're just working with the first column, you could use code like the following:

use strict;
use warnings;

my $line = qq{"229269_2,190594_2,94552_2,266076_2,269628_2,165328_2,99319_2,263339_2,263300_2,99315_2,271509_2,2714",A,1 the next line could look like 84545,X,2};

if ($line =~ /^"(.*?)"|^([^,]*)/) {
    my $column0 = $1 // $2;
    my @nums = split ',', $column0;

    print "@nums\n";
}

The above happens to accomplish the same thing as the previous code. However, it has one big flaw, it's not nearly as obvious to the maintaining programmer what's going on.

Whenever a new coder, or even yourself in 6 months, views the first set of code, it is extremely obvious what format your data is in. You're working with a CSV file, and the first column is a list separated by commas. The second code also works, but the new maintainer must actually read the regex and figure out what's going on to understand both what format the data is in, and whether the code is actually doing it correctly.

Anyway, do whatever you will, but I strongly advise you to use an actual CSV Parser for parsing csv files.

Answer 2

If all you want is all but the last two fields...

   my $string = qq("229269_2,190594_2,94552_2,266076_2,269628_2,165328_2,99319_2,263339_2,263300_2,99315_2,271509_2,2714",A,1);
   $string =~ s/"//g;            # delete the quotes
   my @f = split (/,/, $string); # split on the comma
   pop @f; pop @f;               # jettison the last two columns

   # @f contains what you're looking for