On Sun, Aug 7, 2016 at 12:03 PM, Dominique Pellé
wrote:
> Kevin O'Gorman wrote:
>
> > CREATE INDEX has two problems:
> > 1) poor default location of temporary storage.
> > 2) gets wedged on very large indexes.
> >
> > I'm using the sqlite that came with Xubuntu 14.04,
Kevin O'Gorman wrote:
> CREATE INDEX has two problems:
> 1) poor default location of temporary storage.
> 2) gets wedged on very large indexes.
>
> I'm using the sqlite that came with Xubuntu 14.04, I think it's version
> 3.8.2.
SQLite-3.8.2 is old (Dec 2013). It's better to download and
compile
> Does anybody know where the actual defaults and controlling environment
> variables are documented, by operating system?
https://www.sqlite.org/tempfiles.html
--
Meilleures salutations, Met vriendelijke groeten, Best Regards,
Olivier Mascia (from mobile device), integral.be/om
RETRACTED: Not wedged, just somewhat slower than I expected. I was
expecting about an hour, and it takes 6. On reflection, it's okay. My
ideas about what's going on were wrong too -- anonymous files are used, but
only two of them at a time.
I still think it is using a poor choice of default
Rob,
At 18:27 07/08/2016, you wrote:
Too little sleep and far too much coffee.
I was in the same situation, multiplying by 2 instead of dividing, as
Ryan pointed out.
Nice to see that WAL fits your use case. I for one found it rock solid
and very useful.
--
Jean-Christophe
On Sat, Aug 6, 2016 at 2:49 PM, Kevin O'Gorman
wrote:
> On Sat, Aug 6, 2016 at 2:09 AM, Dan Kennedy wrote:
>
>> On 08/06/2016 09:52 AM, Kevin O'Gorman wrote:
>>
>>> On Fri, Aug 5, 2016 at 2:03 PM, Dan Kennedy
>>> wrote:
>>>
Jean-Christophe
Thanks for the update on wal-mode. Your explanation is clear and makes
sense to us. We can see what we would have a 224MB -wal file, we
experimented with killing processes whilst updating and generally
messing around and SQLite did what is was supposed to do. I wouldn’t
say
Ryan,
Thanks for the update.
We have done a few more tests during the day and not had any issues to
date. This is still on a test version but we are getting a warm, cuddly
feeling about using WAL mode.
The -wal file grows as you describe and you have explained it very well.
We were groping
Andrii Motsok wrote:
> Is that possible to switch database from WAL to DELETE mode when holding more
> than one connection?
No. There is no mechanism to tell the other connections about the change.
> If no, which workaround could we use
Don't do it. What problem do you think you can solve
On 2016/08/07 8:55 AM, Rob Willett wrote:
Richard, Ryan,
Thanks for this. We were dimly aware of WAL but until now hadn’t
needed to use it.
We’ve done a quick check with it and it *seems* to work on a test
database. We’ve all read the docs again and paid attention to
Hi,
Is that possible to switch database from WAL to DELETE mode when holding more
than one connection? If no, which workaround could we use to omit closing of
'all-except-one' connections?
Regards,
Andrii
___
sqlite-users mailing list
Rob,
At 08:55 07/08/2016, you wrote:
We think that using WAL mode works for us, indeed inspection seems to
indicate it does, but the size of the -wal file appears to be far
larger than would be expected. Is there a problem here? It doesn't
appear to be a problem but would welcome any
Richard, Ryan,
Thanks for this. We were dimly aware of WAL but until now hadn’t
needed to use it.
We’ve done a quick check with it and it *seems* to work on a test
database. We’ve all read the docs again and paid attention to
https://www.sqlite.org/wal.html#bigwal
To test if it works we
13 matches
Mail list logo