On Tuesday, 6 August, 2019 04:27, Olivier Mascia inquired:
>On one database instance, a .dump command gives me (among many other
>lines) things like:
>
>INSERT INTO STATISTICS VALUES(11237795927160,11868);
>
>while the output of .recover command gives me things this way:
>
>INSERT INTO
On 6 Aug 2019, at 7:51pm, Dan Kennedy quoted:
>> INSERT INTO 'STATISTICS'('_rowid_', STATDATE, DISKUSED) VALUES( 1,
>> 11237795927160, 11868 );
Quoting these entity names using apostrophes looks wrong to me. It may work
but someone might read it, know it's official output from a program
On 6/8/62 17:26, Olivier Mascia wrote:
On one database instance, a .dump command gives me (among many other lines)
things like:
INSERT INTO STATISTICS VALUES(11237795927160,11868);
while the output of .recover command gives me things this way:
INSERT INTO 'STATISTICS'('_rowid_',
--
The fact that there's a Highway to Hell but only a Stairway to Heaven says a
lot about anticipated traffic volume.
On Tuesday, 6 August, 2019 04:35, test user
wrote:
>So in summary, there is no difference in the multi threaded
>performance that can be gained between SERIALIZED and
On Tuesday, 6 August, 2019 04:35, test user
wrote:
>When SQLITE_THREADSAFE=(1, SERIALIZED), how are the reads/writes
>queued? Is it per connection, file, or process?
>Quote: I don't know exactly what you mean by "how are read/writes
>queued".
>Quote: if you make two simultaneous calls on the
Quote: Please don't try to parallelize a part of your program which exists
mostly to read or write to a SQLite database. SQLite is not a
client/server system. Access to the database file on disk is a
bottleneck. If you have 8 threads which spend most of their time executing
SQLite calls, all
> Le 6 août 2019 à 15:34, David Raymond a écrit :
>
>> "But you do need WAL to achieve multiple readers concurrency..."
>
> Nope, you can have concurrent readers with rollback journal mode. You just
> can't have anyone writing while they're all reading.
>
> (Or I may just be misunderstanding
>"But you do need WAL to achieve multiple readers concurrency..."
Nope, you can have concurrent readers with rollback journal mode. You just
can't have anyone writing while they're all reading.
(Or I may just be misunderstanding what you meant)
___
On 6 Aug 2019, at 11:34am, test user wrote:
> - Does the second request just take longer to return from the FFI call whilst
> waiting for the mutex?
Yes. You do not need to build backoff-and-retry code into your own software.
SQLite does it for you. For every database connection you open,
> Le 6 août 2019 à 14:18, Chris Locke a écrit :
>
>> I got foreign key constraint failures
>
> I don't know why one would work and one would fail, but usually, this
> occurs when you insert a record which has foreign keys to another table,
> but that table hasn't been imported yet. The
> Le 6 août 2019 à 12:34, test user a écrit :
>
> So for example, if I had:
>
> - 8 cores
> - 8 threads
> - 8 db connections, 1 per thread
> - 1 database file
> - x amount of read requests per second
>
> If I were to load balance x requests over each of the 8 threads, all the
> reads would
> I got foreign key constraint failures
I don't know why one would work and one would fail, but usually, this
occurs when you insert a record which has foreign keys to another table,
but that table hasn't been imported yet. The workaround is usually to
ensure all the 'lookup' tables are done
> Le 6 août 2019 à 12:45, Richard Hipp a écrit :
>
> On 8/6/19, Olivier Mascia wrote:
>> Using 3.29.0 with SQLITE_DQS.
>
> Is this the problem that was fixed here:
> https://www.sqlite.org/src/timeline?c=fcd937d9786a82ef
Indeed.
—
Best Regards, Meilleures salutations, Met vriendelijke
On 8/6/19, Olivier Mascia wrote:
> Using 3.29.0 with SQLITE_DQS.
Is this the problem that was fixed here:
https://www.sqlite.org/src/timeline?c=fcd937d9786a82ef
--
D. Richard Hipp
d...@sqlite.org
___
sqlite-users mailing list
Thanks,
When SQLITE_THREADSAFE=(1, SERIALIZED), how are the reads/writes
queued? Is it per connection, file, or process?
Quote: I don't know exactly what you mean by "how are read/writes queued".
Quote: if you make two simultaneous calls on the same connection (from
different threads, for
On one database instance, a .dump command gives me (among many other lines)
things like:
INSERT INTO STATISTICS VALUES(11237795927160,11868);
while the output of .recover command gives me things this way:
INSERT INTO 'STATISTICS'('_rowid_', STATDATE, DISKUSED) VALUES( 1,
> Le 6 août 2019 à 10:27, Olivier Mascia a écrit :
>
> Using 3.29.0 with SQLITE_DQS.
>
> sqlite3 test1.db "create table A(I integer);"
> sqlite3 test2.db "create table B(J integer);"
>
> sqlite3 test1.db
> sqlite> .schema
> CREATE TABLE A(I integer);
>
> sqlite> attach 'test2.db' as cloud;
>
Using 3.29.0 with SQLITE_DQS.
sqlite3 test1.db "create table A(I integer);"
sqlite3 test2.db "create table B(J integer);"
sqlite3 test1.db
sqlite> .schema
CREATE TABLE A(I integer);
sqlite> attach 'test2.db' as cloud;
sqlite> .schema
Error: no such column: cloud
sqlite> detach cloud;
sqlite>
18 matches
Mail list logo