$query->execute() or do { $SQL_FAIL = 1};

$query->execute() or $SQL_FAIL = 1;

This means that you then have to check the value of SQL_FAIL every line before processing the next statement.

Wrapping it all in an eval and using RaiseError really works a treat, looks clean makes your life simpler.

     my $sql_stage=1;
     eval {
          $query->execute();
          $sql_stage = 2;
          $query2->execute();
          $sql_stage = 3;
          $query3->execute();
     };
     if ($@) {
          if ($sql_stage < 3) {
              $fix1->execute();
          } else {
              $fix2->execute();
          }
     }
[download]

If I were going to do something like this, I'd probably do something more like:

$query->execute and $sql_stage++;
$query2->execute and $sql_stage++;
[download]

etc.

That would execute all queries even if one of them failed though. And if after your code finishes, $sql_stage is 1, you still don't know whether it was $query or $query2 that failed. Now add 3 or 4 more queries to that list and the $sql_stage variable becomes practically useless for anything except comparing to the total number of queries.

  $query->execute() || ($SQL_FAIL =1);
  $query->execute() or $SQL_FAIL =1;
[download]