You can do this with a regex:

my @field = m{ \A (.{20}) \s (.{20}) \s (.{3}) \s (.{3}) \z }msx
[download]

But you shouldn’t. Use unpack instead:

my @field = unpack "A20 x A20 x A3 x A3", $_;
[download]

Not least because the A pack template will automatically trim the whitespace for you. To achieve this using a regex, you have to jump through hoops:

my @field = m{ \A (.{1,20}?) \s+ (.{1,20}?) \s+ (.{1,3}?) \s+ (.{1,3}?
+) \z }msx
[download]

Makeshifts last the longest.

Hallo all,

"unpack" is not even mentioned in my "Learn Perl in 24 hours". I need to buy a second book. :-)

THANKS !!

Jean

You probably want to administer The Perl Book Litmus Test. Not that it matters, because Learn $FOO in 24 hours books are always bad.

Makeshifts last the longest.

"unpack" is not even mentioned in my "Learn Perl in 24 hours"

Wow, not mentioned at all? Sheesh, I knew the 24-hours series wasn't great, but wow. Granted, unpack is not one of the most commonly-used functions, and I wouldn't expect a whole chapter on it or anything, but there should be at least a short discription of it. It is, after all, a builtin.

With that said, I'd have used substr rather than unpack for this, personally. *Surely* the book at least has a discussion of substr.

Until you get around to buying a new book, you can get by with perldoc It's not quite the same as having a book in your hands, but it's useful for reference.

If you're looking for a good Perl book, there are some quite excellent ones available. Generally you will not go far wrong with O'Reilly books, for instance. The one with the camel on the cover is great if you have a background in programming in other languages and just need to learn the things that are unique to Perl. If you have little prior programming experience, you might be better off with the llama book. I also quite liked <cite>Effective Perl Programming</cite>, but that one assumes you already know a bit of Perl, so it might not be a good starting point. There are, of course, other good choices as well.

sed -e 's/     /;/g' <YourFile.tsv >OutputFile.ssv
[download]

Update: This assumes consecutive tabs should be converted to the same number of consecutive spaces. Otherwise I think I would be more inclined with:

perl -e 'while( <> ){ s/(\t)+/\;/g; print $_; }' <YourFile.tsv >Output
+File.scsv
[download]

Just a guess based on the sample data, but perhaps what you want to do is export the data using the db tools instead ;-) Or, select using DBI and the appropriate DBD, and then "insert" into a new table using DBD::CSV.

while (<FILE>) {
  chomp;
  print join(";",split (/\s+/,$_)),"\n";
}
[download]

Using split to separate fix-width data sounds like a terrible idea to me. While short data will be padded with whitespace, there's no garantee all data is short - a 6 character entry in a 6 character wide field will not be padded. Furthermore, since the data is fixed-width, there's no need to escape whitespace - there might be whitespace as data. In fact, what you think is padding might actually be part of the data!
Perl --((8:>*

"RESOURCE ORA"

Sod. Didn't notice that. So this is no go.