Hello,
On Mon, Dec 7, 2015 at 5:05 PM, Patrick Donnelly
wrote:
> Update on this:
>
> On Mon, May 5, 2014 at 4:53 PM, Patrick Donnelly
> wrote:
>> Hi,
>>
>> I have an INSERT that looks like
>>
>> INSERT INTO T
>> SELECT ...
>>
>&
On Mon, Dec 7, 2015 at 11:51 PM, Simon Slavin wrote:
>
> On 8 Dec 2015, at 12:19am, Patrick Donnelly wrote:
>
>> There are still writes:
>
> Because you have not defined any transactions, each of your INSERT commands
> it getting wrapped in its own transaction. A tr
On Mon, Dec 7, 2015 at 5:31 PM, Igor Tandetnik wrote:
> On 12/7/2015 5:05 PM, Patrick Donnelly wrote:
>>
>> No rows were inserted but there are several writes. This behavior
>> seems to be caused by AUTOINCREMENT?
>
>
> Could be creating sqlite_sequence table wh
Update on this:
On Mon, May 5, 2014 at 4:53 PM, Patrick Donnelly
wrote:
> Hi,
>
> I have an INSERT that looks like
>
> INSERT INTO T
> SELECT ...
>
> which I'm running numerous times a second that generally does nothing
> because the SELECT returns no rows
a transaction before the next writer, then the
log file will not be restarted? [I assume this is why
SQLITE_CHECKPOINT_TRUNCATE was added?]
[1] https://www.sqlite.org/c3ref/wal_checkpoint_v2.html
[2] https://www.sqlite.org/wal.html
--
Patrick Donnelly
automatic checkpoints are "passive"
and so should not block the checkpointer? I don't see how the writer
is being blocked for so long. Can anyone provide hints on how to
further debug this?
--
Patrick Donnelly
n-standard solutions which
aren't supported in sqlite. For example:
select distinct on (id) id, attribute
from like_this
order by id, random()
from
http://stackoverflow.com/questions/16044828/select-random-row-for-each-group
Any pointers would be appreciated!
--
Patric
to eliminate the INSERT by using a SELECT first to
check if there are no rows? Something like:
CREATE TEMPORARY VIEW V AS
SELECT ...
SELECT COUNT(*) FROM V;
/* If > 0 */
INSERT INTO T SELECT * FROM V;
?
Thanks,
--
Patrick Donnelly
___
sqlite-users mail
8 matches
Mail list logo