>>>>> "Steve" == Steve Ruby <stever@stripped> writes:
Steve> When a user does
Steve> load data infile local
Steve> the file is on the user's computer. The update log doesn't contain the
Steve> data but only the reference to load data infile.
Steve> 1) it is very difficult to figure out which user ran the load data
Steve> 2) the user may have deleted renamed or move the data file
Steve> 3) it is not possible to re-create the data by running the update log
Steve> on the server since the file is on the users's computer and the server
Steve> will be looking for it localy.
Steve> Wouldn't it make more sense if load data put the sql required to insert
Steve> the data into the table in the update log?
I agree! I am just about to start working on some replication issues
and one of these is to make the update log in 'binary' format to make
it smaller and allow one to reliable use the new SQL variables in
sql_statements that updates the table. At the same time I plan to fix
We have of course to introduce some new shell command that will change
the binary update file to a form suitable to feed to 'mysql' and
suitable for reading.
PS: Sorry for the late reply; I am still trying to catch up with the
mails that piled up during my vacation.