Trimming unpack'ed strings

[Anonymous Monk]

Hello monks I'm using unpack to extract fixed width fields from a text file that I need to convert to csv. Because the text in the fields is shorter than the field width I want to trim the whitespace, and there's an added complication of the csv file having fields that aren't in the fixed width file. At the moment I'm doing the most straight forward thing like so (minus the string trimming):

my ($catno, $desc, $pack, $min, $gst, $cost, $unitcost, $retail, $barc
+ode) = unpack("A7 x2 A30 A4 A4 x13 A7 x15 A8 A8 A8 x13 A13", $_);
print CSTORE_OUT join(", ", "", "", $catdesc, "", "", "", "", $catno, 
+$barcode, $desc, $unitcost, "", "", "", $retail, $pack, $gst, "", "")
+."\n";

Is there a one liner I can use to unpack, trim and join? TIA

[almut]

Maybe something like that?

$_ =  "foo        bar        baz     ";
my $catdesc = " extra ";

print join(",", ("", map /^\s*(.*\S)/, unpack("a10 a10 a10", $_), $cat
+desc)[0,0,-1,0,0,3,0,0,1,2,0] );

# output: ",,extra,,,baz,,,foo,bar,"

(I simplified the format, because you didn't provide sample input, and I was too lazy to create a matching string)

That said, writing unnecessarily compact code usually isn't the best idea...

Update: added $catdesc to show that extra variables could in principle also be handled.

[BrowserUk]

If the data are left justified within their fields, then the 'An' template you are using will trim trailing spaces. Unfortunately there is no equivalent for trimming right justified fields.

You could achieve the re-ordering of the fields using the '@n' absolute positioning template. And you can generate the empty fields using 'A0'.

The fly in the ointment as far as producing your output is concerned, it that you have a variable in your output $catdesc, that does not appear to come from your input?

Otherwise you could achieve the one-liner something along the lines of:

print CSTORE_OUT join ',', unpack 
    '(A0)2 @9A30 (A0)4 @0A7 @119A13 @9A30 @90A8 ...', $_;

***NOTE*** That won't work as it's a) incomplete; b) I've re-used the $desc field inplace of the absent $catdesc variable.

It's just an example of what might have been possible had all the output fields come from the input. Whether there is any advantage in doing this as a one-liner is another question.

[gwadej]

I probably would not do this as a one-liner. The trimming can easily be handled with map, though.

my ($catno, $desc, $pack, $min, $gst,
    $cost, $unitcost, $retail, $barcode) =
   map { s/\s+$//; $_ }
   unpack("A7 x2 A30 A4 A4 x13 A7 x15 A8 A8 A8 x13 A13", $_);

The call to map makes a new list from the old list after removing any trailing whitespace.

It seems like there's a way to get unpack to return an empty field. But that escapes me at the moment.

Update: as jwkrahn points out, my change shouldn't be needed because of the 'A'. I should have looked at the unpack template instead of the OP reference to needing to strip the whitespace.

[jwkrahn]

The  map { s/\s+$//; $_ } is superfluous because the  "A" format of unpack already removes all trailing whitespace.