running
vacuum. The developers of sqlite have gone to great lengths to resolve the
vacuum corruption issues, but I tend to error on the side of safety for
production data.
HTH,
Ken
Bambero <[EMAIL PROTECTED]> wrote:
Hi,
I have a small problem with database size. I wrote
Any thoughts on ticket 3007, to disable journalling by passing an omit journal
flag to the sqlite3_open_v2 interface? This would have the I/O load for writes
and probably double the througput.
http://www.sqlite.org/cvstrac/tktview?tn=3007
Thanks,
Ken
You could use a trigger to keep the running total in a seperate table.
"Mahalakshmi.m" <[EMAIL PROTECTED]> wrote: Hi,
I am having 4 records in my Harddisk.
My Processor speed is 400 Mhz.
For "SELECT COUNT(1) FROM MUSIC ;" its getting more time to display the
count.
I have also tried with
ype. Then apply your conversion
function to it!
HTH,
Ken
Liam Healy <[EMAIL PROTECTED]> wrote: I am porting a numerical application from
Oracle to SQLite. For the
most part, I have been successful, but there are slight disagreements
in the floating point number results. I have traced thi
do you have write permissions on mdiadb and the directory???
BanuPrakash G <[EMAIL PROTECTED]> wrote: Database or disk is full is the Error
message i get from the sqlite
:root> ./sqlite3 mediadb.sql
SQLite version 3.5.7
Enter ".help" for instructions
sqlite> CREATE TABLE t1(a INTEGER, b
There can be only one thread that has an exclusive write lock on the database.
Most likely this is a bug in your application.
Heres what I do for my multi threaded app:
1. All transactions use "begin immediate" (that way I don't do a bunch of
work only to find the db locked later).
2. The
I don't see any error handling in your code for SQLITE_BUSY...
Also, Depending upon the odbc driver your using, you might want to change the
begin to a "begin immediate" causing the lock to be taken out at the begin vs
later on...
Xuanvinh Vu <[EMAIL PROTECTED]> wrote:
I did check and I
Try ./sqlite3
BanuPrakash G <[EMAIL PROTECTED]> wrote: Hi,
I have used the latest verision of sqlite i.e. sqlite-3.5.7
Amalgamation release
for PPC Arch.
I was able to cross compile it successfull and when i use the db to
create tables
I see on the SQLITE shell the ERROR message as
se freopen. (why are you re-opening a STDERR in the first place?)
3. Rebuild your logging routines. to open a specific log file.
4. Use syslog ? (if your on a *nix).
HTH,
Ken
Mark Gilbert <[EMAIL PROTECTED]> wrote: Anyone seen this ?
- My app is running happily like
Just give it a try and see what happens. You just need to enable the shared
cache once.
I'd think the blocking would not be any different with the shared cache
enabled. But you should get reduced I/O load since the cache will be larger and
accessible to all threads.
HTH,
Ken
, I have many small db files that have data. These
are "batched" into a db file. If the process fails the in progress
batch will simply be overwritten and re-created. Hence it is an all or
none approach and does not need any recoverability.
Thanks for considering
Just to quantify a cartisian join is a join operation that returns the product
of the rows. So in the example given
episodes 8
foods100
food_episodes 800
episodes e1, foods_episodes fe1, foods f, episodes e2, foods_episodes fe2
Result rows = 8*800*100*8*800 = 4,096,000,000 rows...
It might be helpful to include the version of sqlite.
Have you run your code through a memory analysis routine such as valgrind, to
validate that the leak is not occuring in your application code?
HTH,
Ken
Rob Richardson <[EMAIL PROTECTED]> wrote: Greetings!
I must be doing something
Is there something wrong with the sqlite list? I sent emails to the list over
an hour ago and still nothing is posted?
Thanks,
Ken
___
sqlite-users mailing list
sqlite-users@sqlite.org
http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
directory than the
makefile.in resulted in a cp failure while creating the amalgamated source.
Ken
___
sqlite-users mailing list
sqlite-users@sqlite.org
http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
nd re-created. Hence it is an all or none approach and
does not need any recoverability.
Thanks,
Ken
___
sqlite-users mailing list
sqlite-users@sqlite.org
http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
in the event of a crash.
Thanks,
Ken
___
sqlite-users mailing list
sqlite-users@sqlite.org
http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
How do I prevent the creation of a journal file and a master journal file?
The process does not need recovery as it is an all or none. And is restarted
from the beginning in the event of a crash.
Thanks,
Ken
___
sqlite-users mailing list
sqlite
Do you test for SQLITE_BUSY, when you perform BEGIN IMMEDIATE ?
Yes a process that is reading will continue to read. Once it completes if
anothre process is waiting to write then the additional read locks will not be
granted. This is to prevent writer starvation.
How do you know that the
problem.
HTH,
Ken
Bharath Booshan L <[EMAIL PROTECTED]> wrote:
Hello List,
Here I am stuck with some Database Table corrupt problem and I would
request the people on this list to help me out as I am unable to figure out
the cause for this problem.
Previously I was using SQLite 3.1.3 a
additional
read, to get that data.
So in the end its really your call, Space vs time.
Regards,
Ken
Lee Crain <[EMAIL PROTECTED]> wrote: Understood.
I will be writing our software to search this table by values, so it
sounds like indexing to improve read access is desirable.
The re
succeeds on both systems you may purge the undo data. Or if it fails
you'll need to apply the undo data back on the db's tables.
You could base the undo/redo on:
http://www.sqlite.org/cvstrac/wiki?p=UndoRedo
Good luck.
Ken
Shailesh Madhukar Birari <[EMAIL PROTECTED]> wrote: Hello,
K
or handler. rollback. etc.
hth,
Ken
Tore Austraatt <[EMAIL PROTECTED]> wrote: Ken,
sorry, it didn't make a difference, including _LOCKED in my test.
None of the dropped records are involved in lock situations.
an example...
_exec("BEGIN IMMEDIATE..."
sqlite3
database at the same
time! Sqlite will return a sqlite_busy error.
Check out the following:
http://sqlite.org/lockingv3.html
HTH,
Ken
Tore Austraatt <[EMAIL PROTECTED]> wrote: Thanks, but I'm afraid this don't add
up.
I have tested this in numerous examples. Concurrent INSERT's
disappears
not committing at
exit then a journal file will be left.
HTH,
Ken
Ovidiu Anghelidi <[EMAIL PROTECTED]> wrote: Hi there,
I have a small issue and I was hoping that someone
might be able to point me in the right direction.
I am trying to build a C/C++ application using the
amalgamation v
1. Inserts are not ignored. You can ignore inserts in your app if the db is
busy/locked etc..
2. Test the return code for success/failure.
Tore Austraatt <[EMAIL PROTECTED]> wrote: I find sqlite3 as the perfect choice
for our embedded application - if it
hasn't been for the
issues concering
his stage since B.db is involved in a
transaction, but the journal file is not available for a recovery.
Hope that helps.
Ken
Shailesh Madhukar Birari <[EMAIL PROTECTED]> wrote: Hello,
I want to implement a distributed database using sqlite. Can some one give
me information of whether SQLite
I don't believe that this type of trigger is allowed.
Jeff Hamilton <[EMAIL PROTECTED]> wrote: Hi all,
I have a setup with two databases that have tables which refer to each
other. I'd like to create triggers to handle cleanup when items are
deleted from one database that are referred to from
a prior
prepared statement is still open?
Thanks,
Ken
___
sqlite-users mailing list
sqlite-users@sqlite.org
http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
Running a querry "select count(*) tbl " returned a generic error SQLITE_ERROR.
and an error message of "no such table: tbl"
It would be beneficial If the prepare returned an error code such as
SQLITE_NOTFOUND instead of a generic SQLITE_
ake a look at "BEGIN IMMEDIATE" as it will attempt to lock
at that point. The "begin" does not lock until 1. a cache spill to disk or 2 a
commit.
Colin Darcey <[EMAIL PROTECTED]> wrote: Ken,
I am not sure how to use this facility, looks like one responds by
sending another emai
at the beginning of a transaction.
Ken
Colin Darcey <[EMAIL PROTECTED]> wrote:
I get an error "database is locked" or "database read only" when
attempting to update a table in the database.
I have been reading various articles about this but cannot get
select key where value & '01011';
Try:
select key where (value&176)
you need to convert the bit string to an integer number.
Also be aware that sqlite interperts numbers as signed 64 bit integer and this
can cause some issues with your bit mask number.
HTH.
Jos van
Any reason why the functions above could not be exposed as part of the
sqlite3 api suite?
They sure are handy when writing platform portable code, addressing endian
conversions.
Thanks,
Ken
___
sqlite-users mailing list
sqlite-users@sqlite.org
I'm not 100% sure but
Try splitting your query out into 6 different selects, I think the OR clause
is the problem...
select ID from TableA where column2 like '%test%'
union
select ID from TableA where column4 like '%test%'
HTH
Kalyani Phadke <[EMAIL PROTECTED]> wrote: I am using
In addition to what Igor mentions:
If you want this array to be cross platform portable from Big endian to Little
endian you either need to store some type of endianness flag.
If you dig into the sqlite source you'll find a function like
sqlite3_put_varint.
It will convert 32 bit or 64 bit
I'm curious if the authors of sqlite have given any consideration to the merits
of using a Hash index to retrieve data for In memory Databases ?
Thanks,
Ken
___
sqlite-users mailing list
sqlite-users@sqlite.org
http://sqlite.org:8080/cgi-bin
ORDER BY
Would the above be faster due to the use of the PRIMARY key access ?
Regards,
Ken
Dennis Cote <[EMAIL PROTECTED]> wrote: John Elrick wrote:
> I have a question for the SQLite experts here as to the most efficient
> way to retrieve the following:
>
> Say we have a
Dennis,
Thanks for the info.
Regards,
Ken
Dennis Cote <[EMAIL PROTECTED]> wrote: Ken wrote:
> The following errors out:
>
> CREATE TABLE users (
> user# INTEGER NOT NULL,
> name VARCHAR(30) NOT NULL,
tted
from valid tokens?
Thanks,
Ken
___
sqlite-users mailing list
sqlite-users@sqlite.org
http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
The output of the native sqlite function "hex" is not really a hex value
of a number. Its a hex representation of a string.
select 5629701397680549, hex(5629701397680549) ;
5629701397680549|35363239373031333937363830353439
Where to_hex outputs:
select 5629701397680549, to_hex(5629701397680549)
Try using the attached hex.c function. You'll need to compile this as a .so and
call sqlite3_create_function to register it.
Hth
Mau Liste <[EMAIL PROTECTED]> wrote: Hello all,
I am trying to print some table values in hexadecimal.
I've seen in the docs that there is a function called 'hex'
but
Yes choice is good, Either way I'm greatful for an Excellent Tool, I can
workaround and Live without strict Affinity. But the option to enable it would
be a welcome feature.
Would there be any performance implications of Strict Affinity, either positive
or negative?
Samuel Neff <[EMAIL
have been simpler to debug, if
there had been a type conversion warning or a failure.
Regards,
Ken
Lee Crain <[EMAIL PROTECTED]> wrote: DRH,
">> Can you explain why you think strict affinity mode
>> might be beneficial to you? If somebody can provide a good
>&g
I wrote a function to traverse the linked list and print the zSql to stderr.
There was indeed a sql statement being left open by my code!
Thats a really usefull debugging technique. I might write an extension to
implement this in a more formally.
Thanks,
Ken
[EMAIL PROTECTED] wrote: Ken
The sqlite3_db_handle returns a different address for my statements than the
database session address. (using sqlite3_open)
Should it return the .aDb field instead of the pVfs ?
Thanks,
Ken
___
sqlite-users mailing list
sqlite-users@sqlite.org
and close each statement that has
been opened.
Thanks For any help.
Ken
___
sqlite-users mailing list
sqlite-users@sqlite.org
http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
':
../sqliteSrc/sqlite-3.5.5/sqlite3.c:64663: error: syntax error before numeric
constant
../sqliteSrc/sqlite-3.5.5/sqlite3.c:64700: error: invalid lvalue in assignment
make: *** [sqlite3.lo] Error 1
int isView ; delcaration is wrapped in SQLITE_OMIT_TRIGGER instead of
SQLITE_OMIT_VIEW.
Thanks,
Ken
as
the return code from the Step command.
Thanks,
Ken
sqlite-users@sqlite.org wrote: sqlite-users@sqlite.org wrote:
> Hi all,
>
> Hopefully you can help, while running sqlite version 3.5.5
>
> Two seperate threads each attempt to run a vacuum command against the same
> d
I think this is probably a very simiilar issue as found on the solaris.
The error occurred during asyn3.test.
Here is the gdb output
analyze-99.1... Ok
async3-1.0... Ok
async3-1.1...make: *** [test] Bus error (core dumped)
hermes:herm10:/home/ixion/LIB/sqliteSrc/tst>> ls core
core
,
ppOpen=0x1345a9) at ../sqliteSrc/sqlite-3.5.5/src/os_unix.c:670
Let me know if you need anything else.
Dan <[EMAIL PROTECTED]> wrote:
On Feb 1, 2008, at 6:37 AM, Ken wrote:
> Recompiled without the -O2 flags:
>
> #0 0x00049750 in findLockInfo (fd=4, ppLock=0x1345ad,
&g
In my code I'm getting an SQLITE_BUSY returned from sqlite3_close.
Is there a way to get sqite to tell me which statements it thinks are not
finalized?
Thanks,
Ken
.
Regards,
Ken
Nicolas Williams <[EMAIL PROTECTED]> wrote: I didn't get any core dumps in
Solaris Nevada, but seven tests failed:
Failures on these tests: vtab6-2.2 vtab6-2.4 vtab6-2.5 vtab6-2.6
vtab6-7.1 vtab6-9.1.1 vtab6-9.2
Removed asyn3.test...Since it was core dumping.
tcl version is 8.5.0
gcc is 3.4.2
sqlite 3.5.5
Solairs 8 (5.8)
Test suite errors:
13 errors out of 38318 tests
Failures on these tests: bind-4.4 bind-4.5 printf-1.7.6 printf-1.8.6
printf-1.9.7 tcl-1.6 vtab6-2.2 vtab6-2.4 vtab6-2.5 vtab6-2.6
Recompiled without the -O2 flags:
#0 0x00049750 in findLockInfo (fd=4, ppLock=0x1345ad, ppOpen=0x1345a9)
at ../sqliteSrc/sqlite-3.5.5/src/os_unix.c:670
670 *ppLock = pLock;
*ppLock is Null
Ken <[EMAIL PROTECTED]> wrote: Core/backtrace info:
(gdb) core-file cor
not !=0 but *ppOpen is 0...
James Dennett <[EMAIL PROTECTED]> wrote: > -Original Message-
> From: Ken [mailto:[EMAIL PROTECTED]
> Sent: Thursday, January 31, 2008 2:55 PM
> To: sqlite
> Subject: [sqlite] Solaris bus error
>
>
> After addressing the -l
After addressing the -lrt sched yield on solaris.
make test resulted in:
async3-1.0... Ok
async3-1.1...make: *** [test] Bus Error (core dumped)
Any ideas?
Regards,
Ken
Solaris 5.8 (solars 8)
make test
Undefined first referenced
symbol in file
sched_yield /var/tmp//cckDMcyL.o
ld: fatal: Symbol referencing errors. No output written to .libs/testfixture
collect2: ld returned 1 exit status
unsigned char*)z)[i]] ){
i = 0;
break;
}
}
Maybe your runtime library id different than the source code?
HTH,
Ken
Dennis Cote <[EMAIL PROTECTED]> wrote: Frank van Vugt wrote:
>
> I'll look into it tomorrow, but was under the impression that:
>
> * w
Dennis Cote <[EMAIL PROTECTED]> wrote: Ken wrote:
>
> I think the issue however was that sqlite uses Signed integers And my Number
> although a valid 64bit hex number was to large to fit into a 64 bit signed
> value.
Yes, that's what I said. :-)
> Sqlite simply co
ll proceed with your workaround.
In the mean time I wonder if it would be possible as an enhancement to add
64bit unsigned integer storage into sqlite? And also the capability to convert
strings such as '0x' as this would make writing this type of bit based
logic so much easier.
Regards,
Ken
D
d number?
Thanks,
Ken
I'm not sure if this a bug or if working as intended:
the order by b,a seems to cause B to use asc.. Instead of the desc as written
in the order by statement.
create table tst (a integer, b integer );
sqlite> insert into tst values(1,1);
sqlite> insert into tst values(1,2);
sqlite> insert
assuming your time is in seconds.
try
t1: 201561222 -> 2008-01-28 15:00:22
and you want
t2: 1201561200 -> 2008-01-28 15:00:00
t1 = t2 - (t2 mod 3600)
Joanne Pham <[EMAIL PROTECTED]> wrote: Hi All,
I have a hour table which has the startTime are stored at GMT time. The
Library out of sequence indicates the application is making a call to sqlite in
the wrong order. Recheck your application code and make sure the bind is
against a valid statement handle.
HTH.
Joanne Pham <[EMAIL PROTECTED]> wrote: If I deleted all the database files and
run the program again
Jay
I've used a trigger to do this with good success.
You'll need one trigger per table to cause the delete to cascade through the
tree.
Or if you know that you will always delete from the top level tree"parent"
then just one trigger would probably suffice.
his allows
for a much simpler sqlite_busy/sqlite_locked error handling.
Regards,
Ken
Philip Nick <[EMAIL PROTECTED]> wrote: I tracked down the problem. I was
linking an old version of sqlite3 (3.0.8).
That was the old version we used. I apparently doesn't cache between calls.
My code now performs
or
delete) data.
IPC, is best done as either Message Queues depending upon your data volumes.
Shared memory with appropriate semaphores/ Mutexes.
Or even as one person posted Mmap files. You'll stilll need a mutex or a
semaphore to write.
Ken
"Joshua D. Boyd" <[EMAIL PROTECTED]&g
Sorry if this has been asked, but I'd like to know how to run the test suite.
I could not find any documentation on this on the sqlite website.
Thanks for your help.
Ken
Joe Wilson <[EMAIL PROTECTED]> wrote: --- Ken wrote:
> Doing this in oracle results in an error:
>
> SQL> select max(addr_id), emp_id from z_address;
> select max(addr_id), emp_id from z_address
> *
> ERROR at line 1:
> ORA-00937: not
of
sqlite to the next. Although backwards compatiblity is generally maintained,
there have been instances when this is broken.
I would also maintain in cvs the create sql commands (.schema) seperate from
the data.
Just my opinion, I'm sure others here will differ.
Regards,
Ken
Rich
use. But as a
developer I know better, and write aggregated sql with a group by.
select max(addr_id), emp_id from z_address group by null;
Does not return an error nor does it return data.
Ken
Joe Wilson <[EMAIL PROTECTED]> wrote: --- Darren Duncan wrote:
> At 10:17 PM -0500 1/14/08, Gr
Depending upon your needs an alternative would be to store the numbers list as
a blob. Then the list can be retrieved as a single select.
Just an idea.
Ken
Mike McGonagle <[EMAIL PROTECTED]> wrote: Hello all,
This isn't s question specific to SQLite, so I hope that you don't mind...
but
g pthreads and its detached) then
no harm to the process. But the other threads will never be signalled either,
then they will be stuck.
HTH,
Ken
Andreas Volz <[EMAIL PROTECTED]> wrote: Am Wed, 9 Jan 2008 10:20:31 -0800 (PST)
schrieb Ken:
> Definately use 3.5.4.
>
> Not sure ho
Try:
update zip set city_id = ( select city_id
from city
where name = zip.city);
HTH
Ken
P Kishor <[EMAIL PROTECTED]> wrote: CREATE TABLE zip (zip_code, city);
CREATE TABLE city (city_id, name);
INSERT a
ing useful or
not but just a thought.
0x00 NO_TRACE
0x01 SQL_BEGIN
0x02 SQL_END
0x04 TRIGGER_BEGIN
0x08 TRIGGER_END
Etc...
Regards,
Ken
[EMAIL PROTECTED] wrote: The current interface definition for sqlite3_trace()
is as follows:
void *sqlite3_trace(
s
t, p->db,
"SELECT sql FROM sqlite_master "
"WHERE sql NOT NULL"
" AND type IN ('index','trigger','view')"
" AND tbl_name LIKE shellstatic()"
That should just about take care of it
You might need a few others, to prop
-strip parameter for the .configure as in "-Os" to make
smaller libs a default ???
Thanks,
Ken
[EMAIL PROTECTED] wrote: Ken wrote:
> Dan and DRH,
>
> The amalgamated version as packaged in the link was really nice indeed. Much
> preferred over the full source code
of the make so that loadable extensions are
automatically enabled.
Will future release be of available as almagamated autoconf ?
Thanks,
Ken
Dan <[EMAIL PROTECTED]> wrote:
Hi,
Experimenting with distributing an autoconf configure script
with the amalgamation package. So that it can do the
of the make so that loadable extensions are
automatically enabled.
Will future release be of available as almagamated autoconf ?
Thanks,
Ken
Dan <[EMAIL PROTECTED]> wrote:
Hi,
Experimenting with distributing an autoconf configure script
with the amalgamation package. So that it can do the
(pTemp, db->nextAutovac>=0 ? db->nextAutovac :
sqlite3BtreeGetAutoVacuum(pMain));
#endif
Setting the nextAutovac to a signed char as well resolved this
Regards,
Ken
[EMAIL PROTECTED] wrote: Ken wrote:
>
> AIX had issues. See below.
&g
handle the return codes suche as SQLITE_BUSY, SQLITE_LOCKED
etc.
I don't have any simple examples available, sorry.
Ken
Andreas Volz <[EMAIL PROTECTED]> wrote: Am Wed, 9 Jan 2008 09:03:35 -0800 (PST)
schrieb Ken:
> Andears,
>
> SQLITE 3.5.x is thread safe when configured and compiled
tment. So only way to
exit was cntrl-c (this may be an issue with my system libraries)
If you want me to check anything else please let me know.
Regards,
Ken
Dan <[EMAIL PROTECTED]> wrote:
Hi,
Experimenting with distributing an autoconf configure script
with the amalgamat
Andears,
SQLITE 3.5.x is thread safe when configured and compiled with
--enable-threadsafe.
You can create multiple db connections to a single database.
But only one connection will be allowed to write to the Database at a time.
Take a look at
http://www.sqlite.org/lockingv3.html
ar.
Since SQLITE source code is freely available, you can always create a patch to
the code for the pragma your interested in. And maintain your own version.
Ken
The only occurence of the string "bs" in your code is at the table name
LSOpenJobs
Maybe the DBI layer is not passing everything since there is a carriage return
after the SET?
HTH
Ken
jose isaias cabrera <[EMAIL PROTECTED]> wrote:
Greetings!
I have developed an appli
e
space.
Triggers are another route, but I recommend letting the DB do its job of
storage. Don't add the overhead of adding data scrubbing as well. Do the
scrubbing at the application side where it belongs.
Regards,
Ken
Igor Tandetnik <[EMAIL PROTECTED]> wrote: Ken wrote:
> Some c
lite supports column level check contstraints.
The database is intended for data storage. Not data cleanup. This goes pretty
much for every commercial database out there as well.
Ken
Zbigniew Baniewski <[EMAIL PROTECTED]> wrote: On Sun, Jan 06, 2008 at
07:39:55PM -0800, Darre
Try your test using a "begin Immediate" instead of begin.
A write lock is not taken out until the very last moment (ie a spill to disk
or commit).
Lior Okman <[EMAIL PROTECTED]> wrote:
Hi,
I'm using SQLite3 version 3.4.2 (latest in Debian testing and unstable),
and I have a scenario
t meaning.
SQL> select count(*) from t1 where exists (select x from t1 where x is not
null);
COUNT(*)
--
0
Hope that helps.
Ken
[EMAIL PROTECTED] wrote: The current behavior of SQLite is to not do anything
special
with NULLs in an EXISTS operator. For example:
writing by any other connections.
The dual locking model (prior to the resolution) is ambiguous and could
possibly lead application to deadlocks.
These are just my thoughts on the matter, and are probably not 100% correct.
Ken
Ed Pasma <[EMAIL PROTECTED]> wrote: No, you did not confuse
t prevents other access to the DB.
Ken
Ed Pasma <[EMAIL PROTECTED]> wrote: The ticket has already been resolved, I
see. So it has been
considered a bug. In my earlier reply I tried to defend the current
behavour to be in line with the document, http://sqlite.org/
sharedcache.html. I'm
It may be the case where the index will cause more harm than good in this
instance depending upon the percentage matching. In other words if only 10% of
the records match using the index great... But if say 90% match the index this
would be Very Bad and A Full scan would be quicker.
I full
existing?
And What happens during a crash with two journals ?
This gets complicated very quickly.
Ken
Ed Pasma <[EMAIL PROTECTED]> wrote: Hello,`
Empirically I found that it is exactly true.
Must admit I'm confused but may it is in line with the Shared-Cache
locking model.
This does not m
Continuing with Sams thought process,
Create another table that only contains record id's where value01=value02
Then just query table01_a
The use of an index is probably not going to help and your better off most
likely performing the full table scan especially if you are getting a
Some additional info:
when the sqlite_lock is returned there is another thread that appears to be
reading the same table. Does the sqlite3 step return sqlite_locked in this case?
Thanks,
Ken
Ken <[EMAIL PROTECTED]> wrote:
While using the new 3.5.4 sqlite3_enable_shared_cache
into table... ---> returns SQLITE_LOCKED
Is it possible for both connections to begin exclusive transactions whilst
having the shared cache anabled?
Thanks,
ken
. Then you might be taken more seriously.
As my father was fond of saying; "Money talks and BS walks."
Regards,
Ken
Ion Silvestru <[EMAIL PROTECTED]> wrote: >SW: Richard, We have written to you
directly before to ask you to stop the
>FUD and incorrect statements, and you h
99.9% certain this is not a sqlite issue. But a script language issue.
It depends on how the script language implements threads.
Try writing this in C and using posix threads.
You'll also need to have multiple cpu's to really take advantage of threading
as well.
yahalome
David,
You have some pointer issues going on!
Maybe something like this
struct db_cache_type
{
sqlite3 *db; /* database file pointer */
char *key; /* key identifies the database that was opened */
int mode; /* database open mode */
};
struct db_cache_type **db_cache_list; /* list
201 - 300 of 546 matches
Mail list logo