That's odd. You've not stated the DBI or DBD::Oracle versions.

Try using trace (level 4) with a recent version of the DBI
to see what the DBI thinks is happening to the error status.

Tim.

On Thu, Apr 21, 2005 at 09:03:54AM -0700, Joel Stevenson wrote:
> Hi all,
> 
> I'm running into an interesting problem with a database error that 
> occurs on one statement handle and seems to affect another handle 
> despite using eval {...} and RaiseError.  Perhaps example code would 
> be describe the situation:
> 
> my $attr = {RaiseError => 1};
> my $dbh = DBI->connect ('dbi:mysql:','user','pass', $attr);
> 
> my $topSTH = $dbh->prepare ("SELECT NAME FROM FOO");
> $topSTH->execute ();
> 
> while (my $hashRef = $topSTH->fetchrow_hashref ()) {
> 
>   # do stuff here
> 
>   eval {
>     my $inSTH = $dbh->prepare ("SELECT SOMEVALUE FROM MYMAYBETABLE 
> WHERE ID=1");
>     $inSTH->execute ();
> 
>     my ($value) = $inSTH->fetchrow_array ();
>     # do something with value
>   };
>   if ($@) {
>     print STDERR "ERR: $@";
>   }
> 
> }
> 
> 
> 
> MYMAYBETABLE is a table that may or may not exist and because of that 
> I wrapped the prepare/execute in an eval.  When the table does not 
> exist the error is caught and handled as expected (the "ERR: $@" line 
> is printed to STDERR) but for some reason the next iteration of the 
> while loop throws exactly the same error from the 
> $topSTH->fetchrow_hashref call even though it is not touching 
> MYMAYBETABLE at all.  So I get "DBD::mysql::st fetchrow_hashref 
> failed: Table 'MYMAYBETABLE' doesn't exist at ..." being thrown from 
> both $inSTH and $topSTH !?
> 
> So it seems like the error that occurred in $inSTH is being held onto 
> and affecting other statement handles created from the same database 
> handle.
> 
> I've been looking through the list archives but as yet haven't found 
> anything about this.  Is this expected behavior?  If so, how can I 
> clear the error so that the top while loop can continue on after such 
> an error?
> 
> TIA,
> Joel

Reply via email to