Reading and Writing Hash Records to Text Files (Perl Cookbook, 2nd Edition)

11.10.3. Discussion

If you have an array of records that you'd like to store into and retrieve from a text file, you can use a simple format based on mail headers. The format's simplicity requires that the keys have neither colons nor newlines, and the values not have newlines.

This code writes them out:

foreach $record (@Array_of_Records) {
    for $key (sort keys %$record) {
        print "$key: $record->{$key}\n";
    }
    print "\n";
}

Reading them in is easy, too.

$/ = "";                # paragraph read mode
while (<>) {
    my @fields = split /^([^:]+):\s*/m;
    shift @fields;      # for leading null field
    push(@Array_of_Records, { map /(.*)/, @fields });
}

The split acts upon $_, its default second argument, which contains a full paragraph. The pattern looks for start of line (not just start of record, thanks to the /m) followed by one or more non-colons, followed by a colon and optional whitespace. When split's pattern contains parentheses, these are returned along with the values. The return values placed in @fields are in key-value order, with a leading null field we shift off. The braces in the call to push produce a reference to a new anonymous hash, which we copy @fields into. Since that array was stored in order of the needed key-value pairing, this makes for well-ordered hash contents.

All you're doing is reading and writing a plain text file, so you can use related recipes for additional components. You could use Recipe 7.18 to ensure that you have clean, concurrent access; Recipe 1.18 to store colons and newlines in keys and values; and Recipe 11.3 to store more complex structures.

If you are willing to sacrifice the elegance of a plain textfile for a quick, random-access database of records, use a DBM file, as described in Recipe 11.14.

11.10. Reading and Writing Hash Records to Text Files

11.10.1. Problem

11.10.2. Solution

11.10.3. Discussion

11.10.4. See Also