On 2011/10/20 11:54 AM, Johan De Meersman wrote:
> ----- Original Message -----
>> From: "Alex Schaft"<alexs@stripped>
>>
>> I'm monitoring a mysqldump via stdout, catching the create table
>> commands prior to flushing them to my own text file. Then on the
>> restore side, I'm trying to feed these to mysql via the c api so I can
>> monitor progress (no of lines in the dump file vs no of lines sent to mysql),
>> but the lines are as much as 16k long in the text file times about
>> 110 of those for one huge insert statement.
>>
>> What can I pass to mysqldump to get more sane statement lengths?
> That's a pretty sane statement length, actually. It's a lot more efficient to lock
> the table once, insert a block of records, update the indices once and unlock the table;
> as opposed to doing that for every separate record.
I realize that, I'm just trying to stop the phone calls saying "I
started a restore, and my pc just froze...."
I might just read all the single insert lines, and get a whole lot of
values clauses together before passing it on to get around the
performance issue while having some idea of progress.
Alex
