Bernard,
you could look at the source code of
the mysql client. I think it is written
in C or C++, and should show how you get
the C API working.
Unfortunately there is no global command
to set all sessions to autocommit = 0.
There has been discussion of adding such
an option.
Regards,
Heikki
Ryan,
if you can run the recovery inside gdb
with mysqld compiled with the gcc option -g
and print the stack trace with bt, it may help
me to see where the problem is and fix it.
If you have updated the primary key of a > 8000-byte
row then the table is certainly corrupt and cannot be
saved. But
> Heikki> Hi!
> Heikki> I checked that in the COUNT(*) below InnoDB does
> Heikki> not fetch the whole row if it does the query through
> Heikki> a secondary index. It may sometimes have to
> Heikki> fetch it to determine if the row should be visible
> Heikki> in the current consistent read, the t
Hi!
> "Heikki" == Heikki Tuuri <[EMAIL PROTECTED]> writes:
Heikki> Hi!
Heikki> I checked that in the COUNT(*) below InnoDB does
Heikki> not fetch the whole row if it does the query through
Heikki> a secondary index. It may sometimes have to
Heikki> fetch it to determine if the row should be
> ryc>
> Heikki> Well, mystery solved :). Except that why MySQL uses hours to
> Heikki> sieve away the extraneous rows in the first query.
> >>
> >> To be able to answer the question why:
> >>
> >> SELECT COUNT(*) FROM table1 WHERE col1=0 and col2=0
> >>
> >> Is slow, I would need to get an EXPLA
PROTECTED]>
Cc: [EMAIL PROTECTED] <[EMAIL PROTECTED]>
Date: Thursday, June 28, 2001 1:26 PM
Subject: Re: Re:innobase bug w/ 3.23.29???
>
>Hi!
>
>>>>>> "ryc" == ryc <[EMAIL PROTECTED]> writes:
>
>ryc>
>Heikki> Well, mystery solved
Hi!
> "ryc" == ryc <[EMAIL PROTECTED]> writes:
ryc>
Heikki> Well, mystery solved :). Except that why MySQL uses hours to
Heikki> sieve away the extraneous rows in the first query.
>>
>> To be able to answer the question why:
>>
>> SELECT COUNT(*) FROM table1 WHERE col1=0 and col2=0
>>
> Heikki> Well, mystery solved :). Except that why MySQL uses hours to
> Heikki> sieve away the extraneous rows in the first query.
>
> To be able to answer the question why:
>
> SELECT COUNT(*) FROM table1 WHERE col1=0 and col2=0
>
> Is slow, I would need to get an EXPLAIN for this query.
mysq
Hi!
> "Heikki" == Heikki Tuuri <[EMAIL PROTECTED]> writes:
Heikki> Monty,
Heikki> there is an SQL optimizer deficiency if a two-column index is used.
Heikki> (Added later: actually not, see the end of this email.)
Heikki> Ryan does the following query, for which there is a very good
Heikki>
c <[EMAIL PROTECTED]>
To: Heikki Tuuri <[EMAIL PROTECTED]>
Date: Wednesday, June 27, 2001 12:38 AM
Subject: Re: Re:innobase bug w/ 3.23.29???
>Sorry about the delay in getting you the output. Had a few problems to take
>care of before I could take down the db server for the testin
Ryan,
what does
EXPLAIN select * from tablename WHERE col1=0 and col2=0 limit 100;
show?
What about
EXPLAIN select * from tablename WHERE col1=0 and col2=0;
Have you run
CHECK TABLE tablename;
You have compiled your own mysqld. I can send you a new
mysql/innobase/row/row0sel.c
which w
Paulo,
what MySQL version you are using? I released the source code
for InnoDB on Windows NT and 2000 at my website www.innobase.fi,
but I think it should not work on Win ME, because ME does not support
async i/o. You should compile with
#undef __NT__
on line 18 of mysql/innobase/include/univ
Hi!
I changed it to upcoming 3.23.38 so that it prints an error message
and calls exit(1).
Since you can run out of disk space inside InnoDB only when starting
the database after adding new data or log files to my.cnf, there is
not much need for interaction. If InnoDB would archive its log files
Hi Phil,
>Looked at the Innobase web site - very exciting looking specs.
>SQL:
> -a small interpreter with stored procedure syntax like in
>Oracle, no SQL optimizer
>Does this mean that MySQL using Innobase tables, will this allow for stored
>procedures of
>multiple
14 matches
Mail list logo
