Dear fellow perlmonks,

There is a text-based data format used in the Traveller role-playing game that describes the general characteristics of a star system. The format has a few varieties, but Perl is well-suited to parse it. My question to you is: what methods would you suggest to parse it?

First, the data assumes the following forms:

Old Style       0101 A123456-7 B Lo Po De    A  G
Modern Style    0101 A123456-7 B Lo Po De    A  321 Im G5 III
Extended Style  0101 A123456-7 B Lo Po De    A  321 Im G5 III :0102,01
+03,0104
                ^^^^^^^^^^^^^^
[download]

The "under-careted" fields above are the anchor, the fields which always retain a specific format and width.

All three formats have the same initial fields:

The first group of columns, usually (but not always) 15 characters long, contains the system name. This name can have spaces and nonalpha(\W) characters.

A 4-digit cordinate number is next (\d{4})

A 9-character data string is next (\w{7}-\w).

After 1-2 spaces, there may be an optional character.

After this, a set of codes, all 2+ chars long. These codes can safely be ignored.

Then an optional "travel advisory code", A=Amber, R=Red.

In the "Old Style", a trailing 'G' denotes the presence of a gas giant, while its absence means there ain't one.

In the other styles, there instead are three digits which represent yet more data, followed by a 2-character allegience code and one or more star classifications.

Finally, in the extended style, there may be a colon followed by a comma-separated list of trade route indicators.

In the past, I have used a pair of gargantuan regular expressions to rip out the data; however, I've been thinking there's more elegant ways to deal with it.
For instance, I can take a lot of data out with a match and a subsequent split:

foreach (@row)
{
   my ($name, $rest) = /^(.){15}(.*)$/;
   my @data = split( ' ', $rest, 3 );
}
[download]

But really it's better to just do an initial match:

   my ($name, $loc, $upp, $rest) = /^(\w.*\w)\s*(\d{4})\s*(\S{9}) (.*)
+$/;
[download]

Then the 'rest' can be determined and the data parsed out as necessary:

   my ($code, $pbg, $allegiance, $more) = $rest =~ /(A|R)?\s+(\d\d\d)\
+s+(..)\s+(.*)$/;

or

   my $gg = $rest =~ /G/;
[download]

It's an interesting problem to me, maybe a golf problem, and I'm interested in seeing what folks can come up with.

rje

Comment on Traveller Parsing Select or Download Code