>>>>> "JW" == Jochen Wiedmann <joe@stripped> writes:
JW> I am now asked to change the DBD::mysql driver not to return the result
JW> of mysql_affected_rows(), but read mysql_info() and parse the output to
JW> return the matched rows. This is something I definitely don't like,
JW> because it is slow and unsafe.
JW> My question is whether you can either
JW> - supply the same info via a new function, mysql_matched_rows(), say or
This would be ok, but the DBI interface should expose this function,
not change the meaning of affected rows.
JW> - change mysql_affected_rows to return the number of matched rows
No. This would be bad. I think that the "affected" rows has a valid
meaning in this case, and should not be changed. An SQL server is
well within its rights to rewrite such a query by not affecting rows
that are already set to the "new" value.
Vivek Khera, Ph.D. Khera Communications, Inc.
Internet: khera@stripped Rockville, MD +1-301-545-6996
PGP & MIME spoken here http://www.kciLink.com/home/khera/