On Jan 19, 2011, at 5:35 AM, Richard Hipp wrote:
> On Wed, Jan 19, 2011 at 2:52 AM, Dustin Sallings
> wrote:
>
>>
>> Is it possible without violating any assumptions that would
>> lead to
>> reliability problems to have a DB's WAL exist on a separate
>> filesystem?
>>
>
> No. The WA
Check out:
http://jaksprats.wordpress.com/2010/09/28/introducing-redisql-the-lightning-fast-polyglot/
On Nov 3, 2010, at 10:51 AM, Jay A. Kreibich wrote:
> On Wed, Nov 03, 2010 at 05:10:22PM +0300, Alexey Pechnikov scratched
> on the wall:
>> 2010/11/3 Jay A. Kreibich
>>>
>>> Why not j
create table numbers (val integer);
insert into numbers values (1);
insert into numbers values (5);
sqlite> select * from numbers order by val;
1
5
10
12
12
15
20
20
20
select case when val < 10 then 1 when val >=10 and val < 20 then 2
else 3 end as bin,
count(1) as c
from
Richard,
Would you accept a patch to allow user supplied 'dot' commands in the
shell?
For my own purposes I wanted to have '.load' define some meta commands
as well as custom SQL functions.
Would anyone else find this useful?
Russ
___
sqlite-u
partitions (like above for the web site) is a
great way to scale. The application will know if a table scan is
better or not. Sqlite itself does not have the view.
On Sep 26, 2008, at 1:23 PM, Nicolas Williams
<[EMAIL PROTECTED]> wrote:
> On Fri, Sep 26, 2008 at 12:54:36PM -0400
Perfect solution as long as there is a no index option along with
index by.
On Sep 26, 2008, at 12:54 PM, Russell Leighton <[EMAIL PROTECTED]
> wrote:
> I have another scenario where this is needed , the one in the subject.
> I repeated this problem this AM.
>
> I need
I have another scenario where this is needed , the one in the subject.
I repeated this problem this AM.
I need a 2 key index for some queries and also want to aggregate on
these 2 columns. I need this index BUT I have many large sqlite dbs I
iterate over and they won't fit in the filesystem
A reason I think such functionality would be ideal for sqlite is that
it avoids the need for a fancy query plan optimizer. The user would
have a way to direct the query plan if the simple and obvious plan is
suboptimal.
On Sep 21, 2008, at 11:36 AM, "D. Richard Hipp" <[EMAIL PROTECTED]> wr
Oracle has 'hints' which live in the comments emdedded in the select.
Google 'oracle hint use index'. The 3rd hit down my result list has a
nice overview.(I'd send the link but this stupid iPhone has no cut-n-
paste). I think that hints are really ugly. Not sure about the other
big dmbs.
Alternatively:
pragma planner_ignore_index=1
On Sep 20, 2008, at 11:37 AM, Russell Leighton <[EMAIL PROTECTED]
> wrote:
> It would be very nice to have a way to explicitly control index use.
>
> I'm going to test my theory this weekend but I think if the index is
It would be very nice to have a way to explicitly control index use.
I'm going to test my theory this weekend but I think if the index is
not cached and the data large then the group by is faster without the
index. If this is the case I have a real issue. I need the index for
other queries a
Are there expected performance differences (better or worse) as a
result of the code factoring?
On Aug 30, 2008, at 8:01 PM, D. Richard Hipp wrote:
> SQLite version 3.6.2 is now available on the SQLite website:
> http://www.sqlite.org/download.html
>
> SQLite version 3.6.2 contains rewrites o
I think you are asking about 'table functions', which are functions
that return a rowset and are used in place of a table to generate rows.
See: http://www.postgresql.org/docs/7.3/static/xfunc-tablefunctions.html
To my knowledge this is not supported in sqlite, except perhaps via
virtual tab
On Jul 12, 2008, at 11:08 AM, Dan wrote:
>
> On Jul 12, 2008, at 2:42 AM, Steve Friedman wrote:
>
>>
>>
>> Filip Navara wrote:
>>> how about actually attaching the patch? :)
>>>
>>> - Filip
>>>
>>> On Fri, Jul 11, 2008 at 9:23 PM, Steve Friedman
>>> <[EMAIL PROTECTED]> wrote:
I've just start
On Jun 9, 2008, at 8:52 PM, Igor Tandetnik wrote:
> "Alex Katebi" <[EMAIL PROTECTED]>
> wrote in message
> news:[EMAIL PROTECTED]
>> I am trying to implement remote procedure calls (RPC) for SQLite API
>> to be used in my application.
>> In particular sqlite3_column_double( ) returns a floating p
Digg has an article where it is said that the new Firefox "locks" up
under Linux due to SQLite:
http://digg.com/linux_unix/
Firefox_3_has_system_killing_performance_problem_for_Linux
Bug here:
https://bugzilla.mozilla.org/show_bug.cgi?id=421482
Scanning the bug it seems to b
On Nov 17, 2007, at 4:56 PM, [EMAIL PROTECTED] wrote:
If you compile with -DSQLITE_MEMORY_SIZE= then SQLite
will *never* call malloc(). Instead, it uses a static
array that is bytes in size for all of its memory
needs. You can get by with as little as 100K or so of
memory, though th
On Oct 30, 2007, at 10:18 AM, [EMAIL PROTECTED] wrote:
To accomodate this need, we are considering an incompatible
API change to SQLite. We are thinking of requiring that an
application invoke:
int sqlite3_initialize(...);
I am not sure about the systems that you are trying to support
Could fts3 (the next fts) have the option to override the default
'match' function with one passed in (similar to the tokenizer)?
The reason I ask is then the fts table could be used as smart index
when the tokenizer is
something like bigram, trigram, etc. and the 'match' function computes
If you did not compile sqlite as multi-threaded this is exactly what
would happen.
On Jun 22, 2007, at 9:06 PM, Frederic de la Goublaye wrote:
Hi I just tried this driver:
http://www.zentus.com/sqlitejdbc/
The result is ten times slower or even more.
Maybe I am wrong using this new driver.
Could folks that have used fts2 in production apps/systems relate their
experiences to the group?
I would very much be interested in how folks are using it, how well it
performs with large data and general impressions.
Thanks in advance.
Russ
-
Given you can attach multiple database files in sqlite, perhaps it could
be extended such that:
* When you name an index, you optionally prepend the database name
which can be in another currently attached db
* You attached to multiple db files that have cross references as:
Thx!
[EMAIL PROTECTED] wrote:
Russell Leighton <[EMAIL PROTECTED]> wrote:
So, this was very enlightening...I have a simple backup function that I
now question is correct.
It does:
- execute "begin" // lock from writes
-copy db file to new file byte by byte
-
So, this was very enlightening...I have a simple backup function that I
now question is correct.
It does:
- execute "begin" // lock from writes
-copy db file to new file byte by byte
- execute "commit" // unlock
...I was thinking that "begin" would lock the file.
If I use an flock()
I was afraid of that...it would be cool if someone created a sqlite
server which
handled the networking and serialization...I would take a crack at it
myself but
right now I don't have time.
Dan Kennedy wrote:
I had a musing while reading:
http://news.com.com/2100-7344_3-6063599.html?pa
I had a musing while reading:
http://news.com.com/2100-7344_3-6063599.html?part=rss&tag=6063599&subj=news
where it reminded me of one of MySQL's features:
MySQL's database is built so that it can use a range of different
storage mechanisms, tuned for different purposes, such as
tr
Solved. Bug in my code. Sorry for the alarm.
Russell Leighton wrote:
Update. I actually have 2 nearly identical triggers as described below.
One works, the other produces "shifted" output...it appears that the
sqlite3_column_xxx() functions
are returning the wrong data when ex
produces proper data.
Digging into it further...
Russell Leighton wrote:
This worked in 3.2.8 but not in 3.3.1/2
I delcare a function using sqlite3_create_function()...this takes 1
arg , an integer and calls the usual
code to execute a 'select[ based on the integer as a key and
This worked in 3.2.8 but not in 3.3.1/2
I delcare a function using sqlite3_create_function()...this takes 1 arg
, an integer and calls the usual
code to execute a 'select[ based on the integer as a key and does some
work based on the result.
I declare a temp trigger to call the above wh
(zErrMsg);
I did the same test on a gentoo system with gcc3.3.5 and no issue.
So, I am thinking this is a gcc4.1 bug.
[EMAIL PROTECTED] wrote:
Russell Leighton <[EMAIL PROTECTED]> wrote:
Thx.
I'll re-test with 3.3.1.
That said, isn't 3.2.8 supposed to be the stable relea
Sqlite 3.3.1 passed 'make test' on both of the below platforms that
3.2.8 did not.
I still had the 'date' issues under gentoo...the times missed by
1hr...likely something not quite right in the locale setup.
Russell Leighton wrote:
I get the following failures under
Thx.
I'll re-test with 3.3.1.
That said, isn't 3.2.8 supposed to be the stable release?
[EMAIL PROTECTED] wrote:
Russell Leighton <[EMAIL PROTECTED]> wrote:
Recompiled using 3.2.8 of sqlite, same issue is flagged by valgrind.
3.2.8 is a single-line change to 3
test' under:
FedoraCore4 , gcc4.1
Gentoo 2.6.14-gentoo-r5, gcc 3.3.5
On gentoo I also get these date failures (which I don't see under FedoraCore4):
date-6.1 date-6.4 date-6.5 date-6.8 date-6.13 date-6.16
Russell Leighton wrote:
Recompiled using 3.2.8 of sqlite, same issu
Recompiled using 3.2.8 of sqlite, same issue is flagged by valgrind.
Russell Leighton wrote:
Also, this happens under any constrained insert...from the stack trace
below you would that that would be true. This is confirmed during
another test scenario doing an insert statement into a
Also, this happens under any constrained insert...from the stack trace
below you would that that would be true. This is confirmed during
another test scenario doing an insert statement into a constrained
tabled where I got the same warning about insert.c:980
Russell Leighton wrote
During valgrind ( www.valgrind.org ) testing under linux I was
executing "vaccum" and got:
==17449== Conditional jump or move depends on uninitialised value(s)
==17449==at 0x1CF2200C: sqlite3GenerateConstraintChecks (insert.c:980)
==17449==by 0x1CF233F6: sqlite3Insert (insert.c:629)
=
36 matches
Mail list logo