Re: Speeding up the DBI

[spazm]

The idiom listed above is slightly flawed, perhaps the behavior of fetchall_arrayref has changed in the last few years? Original:

my $rowcache;
while (my $aref = shift(@$rowcache)
|| shift ( @{$rowcache= $sth->fetchall_arrayref(undef, $max_rows)} )
)

From perldoc DBI:

       Here’s an example (assumes RaiseError is enabled):

         my $rows = []; # cache for batches of rows
         while( my $row = ( shift(@$rows) || # get row from cache, or 
+reload cache:
                            shift(@{$rows=$sth->fetchall_arrayref(unde
+f,10_000)||[]}) )
         ) {
           ...
         }

The difference being:  || []. fetchall_arrayref is returning undef instead of an empty arrayref.

The original code as listed will give an annoying "attempt to use undefined value as an ARRAYREF" style error after processing the last line of sql output.

The error line will be attributed to the last line of the while loop since the error happens in the while loop comparator.