On Wed, May 12, 2010 at 11:43 AM, D. Richard Hipp wrote:
> If we drop support for (the non-standard, non-compliant) LinuxThreads
> threading library and instead support only standard Posix threads
> implemented using NPTL, beginning with SQLite release 3.7.0, what
> disruptions might this cause?
>
On 05/13/2010 01:43 AM, ext D. Richard Hipp wrote:
> If we drop support for (the non-standard, non-compliant) LinuxThreads
> threading library and instead support only standard Posix threads
> implemented using NPTL, beginning with SQLite release 3.7.0, what
> disruptions might this cause?
>
>
>I would first create an INTEGER primary key and then place an index on
>name,
>another on i_from, and another on i_to, and then see if the approach below
>has any benefit.
>
>When I tried this with a geo-queryit was actually slower than the standard
>select, and I'm curious if that's always go
* D. Richard Hipp:
> If we drop support for (the non-standard, non-compliant) LinuxThreads
> threading library and instead support only standard Posix threads
> implemented using NPTL, beginning with SQLite release 3.7.0, what
> disruptions might this cause?
There are several variants of NP
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On 05/12/2010 03:23 PM, Pavel Ivanov wrote:
> pthreadMutexNotheld() checks that mutex is not held by current thread
> - if mutex is held by another thread it will return true.
Ooops, I got this completely wrong!
> Although I see that it can return fa
> The mutex is not recursive. However if another thread had acquired the
> mutex during the call to sqlite3Malloc above then the mutex would indeed be
> held by that other thread causing the assertion to fail.
pthreadMutexNotheld() checks that mutex is not held by current thread
- if mutex is hel
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
- From a user backtrace when SQLite is compiled with assertions on, it looks
like an assertion is incorrect. The raw backtrace is later.
pcache1Alloc() contains this code:
pcache1LeaveMutex();
p = sqlite3Malloc(nByte);
pcache1EnterMutex(
I would first create an INTEGER primary key and then place an index on name,
another on i_from, and another on i_to, and then see if the approach below
has any benefit.
When I tried this with a geo-queryit was actually slower than the standard
select, and I'm curious if that's always going to be t
On Wed, May 12, 2010 at 10:40 AM, Roger Binns wrote:
> On 05/11/2010 11:50 AM, Matt Young wrote:
>> sqlite> create virtual table if not exists words using fts3 (f1 );
>> Error: near "not": syntax error
>
> http://www.sqlite.org/cvstrac/tktview?tn=2604
>
> To fix it requires code changes to SQLit
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On 05/11/2010 11:50 AM, Matt Young wrote:
> sqlite> create virtual table if not exists words using fts3 (f1 );
> Error: near "not": syntax error
http://www.sqlite.org/cvstrac/tktview?tn=2604
To fix it requires code changes to SQLite and the SQLite
Thanks. A novice sometimes posts prior to a complete search of documentation.
On 5/11/10, P Kishor wrote:
> On Tue, May 11, 2010 at 1:50 PM, Matt Young wrote:
>> sqlite> create virtual table if not exists words using fts3 (f1 );
>> Error: near "not": syntax error
>> sqlite> create table if no
Hi,
Given the following table with large row count 'row_count':
CREATE TABLE table
(
i_name TEXT,
i_from INTEGER,
i_toINTEGER,
i_data BLOB
)
I am wondering what would be the fastest way to get all rows with a
given name 'myname' that inte
If we drop support for (the non-standard, non-compliant) LinuxThreads
threading library and instead support only standard Posix threads
implemented using NPTL, beginning with SQLite release 3.7.0, what
disruptions might this cause?
Is anybody still using LinuxThreads?
D. Richard Hipp
d...@h
On Wed, May 12, 2010 at 4:18 AM, Gerhard Häring wrote:
> A pysqlite user reported a problem with COALESCE. I could find out
> that his test case worked ok with SQLite versions 3.6.12 through
> 3.6.20, but failed with 3.6.21 to 3.6.23.
>
The problem was actually introduced in 3.6.12. Here is a s
On Wed, May 12, 2010 at 06:23:46AM -0500, Black, Michael (IS) scratched on the
wall:
> Somebody please correct me if I'm wrong on this one...
> I think sqlite can work with multiple processes just doing read-onliy
> operations (like SELECT).
Yes.
> It's just the write operations
> (INSERT/UP
You haven't said what kind of machine you're running on.
I did this test using 9,000,000 records and got 40,000 inserts per second.
sqlite3 test.db :
> On Tue, May 11, 2010 at 12:47 AM, ?? wrote:
>> Dear
>>
>> I have about 9 million data insert string need to insert into an table ,each
>> row
On Tue, May 11, 2010 at 11:28:14PM -0700, Matt Young scratched on the wall:
> C:\R\SQLite>sql ap.data.3.Food.db
> SQLite version 3.6.23.1
> Enter ".help" for instructions
> Enter SQL statements terminated with a ";"
> sqlite> .tables
> Col2U seriesdata seriesid
> sqlite> .q
>
>
H...I duplicated this behavior on 3.6.23.1...I can only assume this is by
intent as the interpreter isn't designed for mulitiple attached databases.
.dump doesn't work on an attached table either
Michael D. Black
Senior Scientist
Northrop Grumman Mission Systems
SQLITE_BUSY is not an error...just a fact.
All your processes cannot work on the database at the same time...at least not
when one of them is doing an insert. You could be changing the table while
you're scanning it. EXXCLUSIVE doesn't change that idea.
Somebody please correct me if I'm wro
A pysqlite user reported a problem with COALESCE. I could find out
that his test case worked ok with SQLite versions 3.6.12 through
3.6.20, but failed with 3.6.21 to 3.6.23.
I could narrow down his test case to this very simple one below. The
combination of addition, named parameters and COALESCE
On Tue, 11 May 2010 13:05:10 + (GMT), kundan bharti
wrote:
>Hello Sir/Maam,
>
> I am getting the error "SQLITE_CORRUPT: database disk image
> is malformed" while opening the attached database in
> SQlite Expert.
Attachments usually don't make it in mailing lists, perhaps
you can put it on a
Emany,
this has possibly already been done, but have you got an index on the table
for the column called "description"?
I would suggest a timed test on both the Symbol device and the computer
before and after the index is added.
Also, the Symbol may only have 64Mb of RAM, which probably would sl
22 matches
Mail list logo