unlink single line of data

[envar]

This is a project that has been back-burner'd - and now I'm faced with adding an unlink (delete record) routine.

Data is stored in a flat file database using an 'old' style Perl cgi. Each record consists of many fields. Each record is written to two separate databases - one db per category (category.db) and one that stores all category's records (big.db). Each db has anywhere from 50 to 100 record lines.

a record line may look like:

state:NM|food:chili|color:blue, purple, red green|etc:etc| (lots of fields that end in a pipe)

Using a form to select which record to edit (delete) - how do I unlink that "line" from the category.db and big.db? I don't know how to get all those fields into an array and unlink.

I can unlink a particular file easy enough - but I've never had to unlink a line from a data file, let alone two different data files at the same time.

Sorry, but I really need some pointers on this. It's down to the wire with this project.

 

[yauncin]

Something like:

open FILE, '/path/db';
@slurp = <FILE>;
close FILE;
open FILE, '>/path/db';
foreach (@slurp)
{
next if #(condition to find line you want removed);
print FILE;
}
close FILE;

This will take all the lines from your database file and write them back out unless the line matches some criteria (By way of some regex). I would suggest writting to a test file until you're satisfied your script works.

 

[Coderifous]

Or use the splice() function. Yauncin's method above is actually a pretty good one, except that with a large file your using alot of memory, another way to do it would be:

open INFILE, '/path/db';
open OUTFILE, '/path/newdb';

while <FILE> {
print OUTFILE unless #insert matching pattern here;
}

close(INFILE);
close(OUTFILE);

rename '/path/db','path/olddb';
rename '/path/newdb','/path/db';

This method saves on memory a bit, and gives you a backup file.

Hope this helps

--Jim