Help with subroutine

ssmith001 · #1 (**permalink**) 09-21-06, 14:41

I have the following subroutine that 99.99% of the time will return a single value. My concern the how do I build in error trapping for the .01% in case there happens to be more than two distinct values. I want to bail.

Code:

sub what_category_do_we_process
{
  my $planner = shift;
  my ( $cur, $sql );
		
  $sql = "SELECT DISTINCT category 
  				  from promo_params 
  				 WHERE plnrcode = ? 
  				   AND processflag = 'N'";
               
  $cur = $dbh->prepare($sql);
    
  $cur->execute($planner);
	
	$cur->bind_columns(undef, \$category);

	while($cur->fetch()) 
	{	
   return $category;
	} 

}

Peter.Vanroose · #2 (**permalink**) 10-02-06, 08:43

What about:

Code:

sub what_category_do_we_process
{
  my $planner = shift;
  my $sql = "SELECT DISTINCT category 
		  FROM   promo_params 
  		  WHERE  plnrcode = ? 
  		    AND  processflag = 'N'";
               
  my $cur = $dbh->prepare($sql);
  $cur->execute($planner);
  $cur->bind_columns(undef, \$category);
  $cur->fetch();
  my $cat = $category;
  if ($cur->fetch()) { die "Multiple rows returned"; }
  else { return $cat; }
}

sco08y · #3 (**permalink**) 10-07-06, 20:09

Just to add to what Peter posted, even if you're going to bail, you might need to clean up database connections or something.

Code:

eval {
   blah;
   blah;
   do_stuff_with(what_category_do_we_process($blah));
   # 99.9% of the time it just works, and continues here...
   blah;
   blah;
}
if($@ eq "Multiple rows returned") {
   handle_multiple_row_exception;
} elsif($@) {
   die $@;
}

See perldoc perlvar for an explanation of $@ and perldoc -f eval for more info.

boftx · #4 (**permalink**) 12-02-06, 12:49

How about this? You can wrap the call to "what_categroy_do_we_process" in and eval as suggested before to trap the die when that happens.

Jim

Code:

sub what_category_do_we_process
{
  my $planner = shift;
	
  my $sql = "SELECT DISTINCT category 
    from promo_params 
    WHERE plnrcode = ? 
    AND processflag = 'N'";

  my $cats = $dbh->selectcol_arrayref($sql,undef,$planner);

  die "More that one category found" if ( @{$cats} > 1 );
  die "No categories found" unless ( @{$cats} );

  return($cats->[0]);
}

sco08y · #5 (**permalink**) 12-03-06, 22:05

Quote:

How about this?

Yup. Now, if you are doing a lot of stuff that requires all this cleanup, you might want to reuse your cleanup code. Here's how:

Code:

sub errorcheck_db_action(&) {
    my($func) = @_;
    eval {
        &$func();
    }
    if($@) {
        if($@ eq "Multiple rows returned") {
            whatever_cleanup();
        } elsif(...) {
        ...
        } else {
            die $@; # if we don't know what kind of exception it is, pass it along
        }
    }
}

errorcheck_db_action {
    do_database_stuff();
};

If you're confused as to what sub foo(&) { ... } does, check perldoc perlsub for prototypes. Basically, the stuff inside the curlies becomes an anonymous subroutine without the "sub" keyword. It even gets its own @_ and everything, and if you have values outside it, it sees them. (It's a thing computer scientists call closures... it just works.)

All you need to iron out is where you're going to stash your DBI object. If you subclass DBI and use it as a method, you lose the prototype, so calling it would look like $dbi->errorcheck_action(sub { block_of_code; }). The syntax wouldn't be as nice, but it'd work.

Internet Services

Web Development

Digital Marketing

Consumer Tech