With the testing data I am using, the file is generated with no obvious problems at the 17985 row mark, but if 17986 rows of data are used, the file is sent (file is generated to STDOUT by an Apache CGI script) as a zero byte file.

Each row is 14 columns wide (the result of a fairly ugly query), the boundary region falls far short of the 65536 row count for excel sheets, the largest working files are not too far shy of the 7MB limit I have seen associated with excel files, and there doesn't seem to be any problems with the actual data that is being used (no weird characters, the break line changes depending on previous lines, ie if I throw away the first line of data, the break zone moves down), so I assume the problem is related to the 7MB limit.

Unfortunately Spreadsheet::WriteExcel::Big isn't playing nicely - it won't generate a usable file with even the first 10 rows of the 17000+ rows of data that were successfully used with standard Spreadsheet::WriteExcel. I have upgraded from 2.11 to 2.15 of S::WE to try and eliminate any bugs that may have been present and subsequently fixed, but the problem still persists.

The relevant code fragment is:

require Spreadsheet::WriteExcel::Big;  #1A
print $foo->header(-type => 'application/vnd.ms-excel', -attachment =>
+ 'export.xls', -expires => 'now');
binmode(STDOUT);
my $export_sheet = Spreadsheet::WriteExcel::Big->new(\*STDOUT) or warn
+("Unable to create new Excel document"), do_empty(), exit;  #1B

my $sheet = $export_sheet->add_worksheet('Support Export');

##spreadsheet header row and formatting elided##

my $row_num=1;

my $stop = 17986;  #2
$stop = 10;  #3
my $count = 0;
#$results is an arrayref of hashrefs from a db query
foreach(@$results)
{
 $count++;
 last if $count > $stop;
 $sheet->write_row($row_num++, 0, [$_->{issue_id}, $_->{issue_name}, $
+_->{contact_name}, $_->{company_name}, $_->{staff_name}, $_->{product
+_name}, $_->{issue_status}, $_->{issue_submitted}, $_->{issue_opened}
+, $_->{issue_resolved}, $_->{issue_closed}, $_->{billable}, $_->{non_
+billable}, $_->{issue_resolution}],);
}
$export_sheet->close;
[download]

As it is, the generated file is zero bytes. If the ::Big is removed from lines 1A and 1B this will generate a small version of the file, as expected. Removing line 3 will result in the file being zero bytes again. Altering line 2 to 17985 leads to a file of size 6.81 MB (7,143,936 bytes). (these steps are performed in that order)

Q 1)Am I missing something obvious with ::Big? I am following the instructions as I understand them.

Q 2)If, as I suspect, this is simply due to the file not generating once it hits the magic threshold, is there an easy way to handle this? If necessary I could get away with truncating (but I would obviously have to know where to truncate), and giving a message to use more explicit search criteria, but giving the full data set would be preferable.