rm bit shifting and will be much faster than the
> string conversions.
You're thinking of the varint format used in the file-format:
http://www.sqlite.org/fileformat.html#varint_format
sqlite3_int64 is just a typdef of int64_t on most platforms.
Dan.
>
>
> HTH,
&
ement?
In this case it will only matter if you have statements that affect more
than one row that could hit a constraint. i.e. an UPDATE or an INSERT
INTO
... SELECT statement.
Dan.
___
sqlite-users mailing list
sqlite-users@sqlite.org
http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
; INTO
>> ... SELECT statement.
>
> So, if I always do INSERT ... VALUES (), i.e. always insert only one
> row by one statement, then this insert will be able to rollback
> harmlessly even without journal?
Assuming it's a constraint failure that causes the rollback that
hed through.
> And another question just to be sure that I understand it correctly:
> async vfs holds only one queue for all opened database files, right?
True statement.
Dan.
___
sqlite-users mailing list
sqlite-users@sqlite.org
http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
torSize() and
xDeviceCharacteristics().
But that's all, nothing is read, written or locked until later on.
So you won't get SQLITE_BUSY back from sqlite3_open_v2(). You might
manage
an SQLITE_CANTOPEN if you tried to open a directory or, depending on the
flags passed t
by a power failure or OS crash. When using asynchronous IO this should
not
be possible (assuming the hardware is not being untruthful - just as
when
using regular "PRAGMA synchronous=full" mode without the async IO VFS).
Dan.
___
sqlite-users mailing list
sqlite-users@sqlite.org
http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
t; same
> operation on the DB multiple times, after a while everything will
> become
> slow.
Maybe try this patch:
http://www.sqlite.org/cvstrac/chngview?cn=6619
Dan.
___
sqlite-users mailing list
sqlite-users@sqlite.org
http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
On May 14, 2009, at 4:08 AM, Ofir Neuman wrote:
> Thanks Dan this was very helpful it seems that I have a similar
> problem.
>
> Didn't try the patch yet but I find out a scenario to reproduce the
> problem.
>
> 1. Process A perform UPDATE on bunch of record
On May 14, 2009, at 2:33 PM, Ofir Neuman wrote:
> Dan,
>
> Before I got your reply I checked it with Pcache1RemoveFromHash()
> and it
> fixed the problem.
>
> I will re-check this issue with pCache->nPage-- but I guess that it
> will fix
> the problem as
enough to figure out why the pointer is decremented, but it seems this
> is causing core dumps only with virtual tables.
>
> Did anybody else see this? Am I doing something wrong?
Do you have a stack trace for the crash? On linux, open the core with
[gdb ] and type "where".
Dan.
r threads to do both writes and reads - bad
> for
> performance.
Using the same thread/connection-handle for reading and writing a
database
might improve performance.
Dan.
___
sqlite-users mailing list
sqlite-users@sqlite.org
http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
/sbin:/bin:/usr/
> games
> So I think the executable file sqlite3 is seen by PATH.
> Could you please tell me any more suggestions?
$ ldd /usr/local/bin/sqlite3
Check it's linking to the correct libsqlite3.so.
Dan.
>
>
>
>
>
>
> _
to table A;
> (2) select a row from B ,then insert the row into table A
> immediately, repeat the precess.
Usually strategy 2. Falls back to 1 if A and B are the same table, or
if there are triggers that imply inserting rows into B may modify A.
Dan.
>
>
>
>
>
> __
e has
> finished.
Did you open the flash database with "PRAGMA locking_mode=exclusive", or
some other option that would turn on persistent journals?
Dan.
___
sqlite-users mailing list
sqlite-users@sqlite.org
http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
from that table until connection A either COMMITs or
ROLLBACKs
its open transaction. Trying to do so returns SQLITE_LOCKED. However,
when
in read_uncommitted=1 mode, the second connection reads the modified
contents
of the table, even though that data has not yet been (and indeed may
never be
On May 25, 2009, at 2:52 PM, Manasi Save wrote:
> Hi,
>
> Can anyone help me on how to run sqlite3-3.6.14.1.bin.gz on Linux
> machine.
> I am working on Fedora 10. Please if anyone can provide any input on
> this.
Like this:
d...@computer2:~/tmp$ wget http://www.sqlite.org/sqlite3-3.6.1
een such
> documents and reality should always be reported, so here it is.
As you say, probably not important in practice but still worth getting
right. Thanks for reporting this.
Dan.
___
sqlite-users mailing list
sqlite-users@sqlite.org
http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
>
Thanks very much for catching these. Document is now updated with the
fixes in the fossil docsrc repository.
Dan.
> 1. In the following, s/less than/less than or equal to/
>
> """
> 2.3.3.4 Index B-Tree Cell Format
> [snip 2 paragraphs]
>
> If t
Looks like the same issue as reported here:
http://www.sqlite.org/cvstrac/tktview?tn=3869
Fix will be included in 3.6.15. Or just grab the patch here if you're
in a hurry:
http://www.sqlite.org/cvstrac/chngview?cn=6666
Dan.
On May 27, 2009, at 12:12 PM, Brian Dantes wrote:
&g
27;m curious what
> is the benefit or application of a compound index compared to two
> single indexes ?
Such a query can use the compound index T1Idx above. It can do the
following
too:
WHERE b = 3 AND c = 4;
WHERE b = 3 AND c > 4;
but cannot be used to optimize:
WHERE c =
On May 27, 2009, at 10:13 PM, Marcus Grimm wrote:
> Hi Dan,
>
> thanks for your answers: this was exactly what I was looking for.
>
> Dan wrote:
>> On May 27, 2009, at 9:08 PM, Marcus Grimm wrote:
>>>
>>> The difference, I guess, to a "real" sq
On Jun 10, 2009, at 5:23 AM, Joanne Pham wrote:
> Sorry Couldn't locate the email about Compite with
> DSQLITE_THREADSAFE=1 bu the application has multiple threads using
> the same connection?
> Would you pleas direct me to any document that has this info.
> Thanks,
> JP
>
http://www.sqli
_reset()
or sqlite3_finalize() on the statement handle.
You can also invalidate it by calling sqlite3_column_text16(), or
bytes16().
In this case the buffer that contains the utf-8 representation (a
pointer
to which is returned by sqlite3_column_text()) is discarded to make room
for the utf-16
piled into the app, and seem to
> stand apart from the core sqlite3, I was curious if I could use them
> with sqlite3 version 3.5.9 since that version is already in use in
> my organization.
Might very well work on unix systems. I think there were some
incompatibilities
with the windows
e
the database was unlocked and modify table t2, changing the set of rows
selects.
Of course, if the asynchronous thread manages to commit the first
transaction
to disk before the user has time to execute the second, the database
file
will be unlocked between transactions.
Dan.
__
On Jun 20, 2009, at 6:22 AM, Robert Lehr wrote:
> Once again, I am forced to reply to the top-level, outside the thread
> b/c I am not receiving individual replies. I apologize for the
> inconvenience.
>
> On Jun 20, 2009, at 12:30 AM, Dan wrote:
>> On Jun 20, 2009, at
ment handle should be sqlite3_reset() or
sqlite3_finalize().
Dan.
> This
> way I could try to interrupt the statement many times so I can
> continue to process the main loop to get the UI going.
>
> Thanks.
> ___
> sqlite-users m
On Jun 26, 2009, at 11:13 PM, Shaun Seckman (Firaxis) wrote:
> I have a column ("ID") in a table that is the primary key integer so
> it
> should be an alias for ROWID. Is it safe to have a ROWID of 0?
Yes.
>
>
>
>
> -Shaun
>
>
>
>
>
> ___
> sqli
>fprintf(stderr, "sqlite3_stmt: %s", sqlite3_errmsg(db));
>goto out2;
>}
>
>ret = sqlite3_bind_text(stmt, 1, "1.1", strlen("1.1") + 1,
>SQLITE_TRANSIENT);
>if (ret != SQLITE_OK) {
>fprintf(stder
On Jun 29, 2009, at 2:44 PM, Kris Groves wrote:
> Given the trigger creation statement :
> CREATE TRIGGER triggername DELETE ON tablename BEGIN statements; END
>
> When does it happen ? BEFORE or AFTER ?
>
> I've read http://sqlite.org/lang_createtrigger.html, but it is
> unclear to me. If I h
On Jul 4, 2009, at 8:12 PM, Wenton Thomas wrote:
> I use prepare statements, and I am sure I finalize all of them.
You may be mistaken. sqlite3_next_stmt() will return 0 if all
statements really are finalized.
assert( sqlite3_next_stmt(db, 0)==0 );
e system-installed version. :-)
You will need to build from the source tarball (the one with the
unsupported configure script), not the amalgamation package for
this to work.
Dan.
> Can anyone help me?
>
> Thanks,
>
> Andy
>
>
>
>
> ___
le and include RTREE and UPDATE_DELETE_LIMIT.
>
> Any ideas? I tried putting a
>
> #define SQLITE_ENABLE_UPDATE_DELETE_LIMIT 1
>
> into sqlite3.h, but that didn't work. :(
The makefiles in the full source tarball support a target "sqlite3.c".
Build this targe
elect query
> with "column_name LIKE 'a%' ".
> So my first question is will this query use the index created or not?
Possibly. Details are here:
http://www.sqlite.org/optoverview.html#like_opt
Dan.
___
sqlite-users mailing li
On Jul 13, 2009, at 3:38 PM, Wilfried Mestdagh wrote:
> Hi,
>
> It seems that if I create the table as a field type 'char' instead of
> 'string' then the issue is solved. But I thought the field type was
> of non
> importance?
Details here:
http://www.sqlite.org/datatype3.html#affinity
>
http://www.sqlite.org/cvstrac/getfile?f=sqlite/ext/async/sqlite3async.c&v=1.7
Dan.
___
sqlite-users mailing list
sqlite-users@sqlite.org
http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
xSize I don't think will be
> possible to implement at all.
Write a wrapper that allocates the requested size plus an 8-byte header
each time an allocation is requested. Use the header space to store the
size of the allocation.
Dan.
> ___
&g
On Jul 23, 2009, at 10:50 PM, Jay A. Kreibich wrote:
> On Thu, Jul 23, 2009 at 09:46:24AM -0400, Wilson, Ron P scratched on
> the wall:
>
>> I think the OP just has the wrong expectations.
>
> Yes, and no. While SQLite doesn't have a one-click-to-install
> download, I have to agree that the
e lock, delete records from the database
file, then downgrade back to a read lock.
Dan.
>
> As a second note: it's pretty bad idea to change table which is used
> in not yet finished select statement. In SQLite it can lead to
> unpredictable behavior. E.g. rows that you've alr
While compiling any statement (sqlite3_prepare_v2()), or while stepping
(sqlite3_step()) a statement that accesses the main database, a mutex
associated with the in-memory cache of the main database will be held.
Dan.
On Jul 27, 2009, at 11:16 PM, Alessandro Merolli wrote:
>
atabase doesn't match that of the machine
it is used on. But not much.
Dan.
-
To unsubscribe, send email to [EMAIL PROTECTED]
-
t transactional? VACUUM is obviously
one of them, are there any other?
Some of the pragma statements. Can't think of anything else.
Dan.
-
To unsubscribe, send email to [EMAIL PROTECTED]
-
with everything else.
Dan.
Mike
-Ursprüngliche Nachricht-
Von: Dan [mailto:[EMAIL PROTECTED]
Gesendet: Dienstag, 27. November 2007 15:59
An: sqlite-users@sqlite.org
Betreff: Re: [sqlite] Transactional DDL
On Nov 27, 2007, at 9:47 PM, Igor Sereda wrote:
I noticed that CREATE TABLE
n built into it. Use testfixture to run the script
"quick.test"
from the test directory (takes from 2-4 minutes). Once that passes, run
"all.test" (warning - takes upwards of 30 minutes to run).
$ ./testfixture ../sqlite
ith normal operation. Not
100% sure, just guessing...
Dan.
Thanks!
Mark
--
---
To unsubscribe, send
b FROM x2 ORDER BY [x1.b];
SELECT x1.b, a FROM x1 UNION SELECT a, b FROM x2 ORDER BY 1;
Because it is consistent with this kind of statement:
SELECT "x1.b" FROM (SELECT x1.b FROM x1);
Any opinions?
Dan
#x27;s internal schema until step 6. By the time it
is loaded for the first time in step 6 the "student_master" table has
already been added. Hence no problem.
Dan.
-
To unsubscribe, send email to [EMAIL PROTECTED]
-
On Dec 4, 2007, at 10:35 PM, Joe Wilson wrote:
--- Dan <[EMAIL PROTECTED]> wrote:
i.e., if we have:
CREATE TABLE x1(a, b, c);
CREATE TABLE x2(a, b, c);
then the following pairs of statements are equivalent:
...
SELECT x1.b, a FROM x1 UNION SELECT a, b FROM x2 ORDE
On Dec 4, 2007, at 11:47 PM, Dennis Cote wrote:
Dan wrote:
This is failing because the internal representation of the database
schema used by the first connection has not yet been updated to
include the changes made in step 4 by the second connection.
After the sqlite3_prepare() in step 6
On Dec 5, 2007, at 1:24 AM, Joe Wilson wrote:
--- Dan <[EMAIL PROTECTED]> wrote:
The "b" in the ORDER BY does not match "x1.b" because it is
not a simple identifier (according to matchOrderbyToColumn()).
It does not match either "" or " as ".
When issuing PRAGMAS, do they have to be issued from sqlite executable
or can they also be issued via the Tcl API?
Pragmas can be executed via the Tcl API, because they are a part of
SQLite's SQL dialect. Unlike the 'dot-commands' that are only available
in the s
7;2008-01-31 23:59:59.99';
sqlite> .quit
In English, why does adding the 'time' portion to the between
clause not
find the record?
Because according to the default collation sequence, strcmp(), the
string '2008-01-01 00:00:00' is larger than '20
On 10/12/2007, Kees Nuyt <[EMAIL PROTECTED]> wrote:
On Mon, 10 Dec 2007 12:22:53 +0100, Clodo <[EMAIL PROTECTED]>
wrote:
In this case:
-
CREATE TABLE test (
Field01 text PRIMARY KEY NOT NULL,
Field02 text
);
insert into test values ('alpha','beta');
update test
/www.sqlite.org/pragma.html)
help any?
Dan.
(a) Since I am new to SQLite, I would like to hear from the
community on how
this can be done using what SQLite3 provides.
If there are specifics that need to be managed outside the context of
SQLite, I am fine with that. However, for doing extern
trace in gdb I gather the
following
info:
Does the sqlite3 shell work when compiled the same way?
Dan.
-
To unsubscribe, send email to [EMAIL PROTECTED]
-
statements are being executed, I would like to - for
example
- redirect the queries to another data base when 70% of the space is
reached. How can I do that?
I think you would have to check the size of the file on disk using
external (non-SQLite) APIs before each insert.
Dan.
~Yuva
On Dec
roblem?
Should be fixed by:
http://www.sqlite.org/cvstrac/chngview?cn=4643
Please post if you find this is not the case (or if you find it is).
Dan.
-
To unsubscribe, send email to [EMAIL PROTECTED]
-
retrieve a value, it is truncated to 32-bits before returning it.
So to be safe, you're better off always using sqlite3_column_int64().
Dan.
I don't know whether or not I'm just "lucky" that the application
which wrote the database uses 64 bit for all of its integers?
his out post if it fails
to install, fails to detect readline, fails to build shared
library on your AIX system, that sort of thing. Thanks in
advance.
Dan.
-
To unsubscribe, send email to [EMAIL PROTECTED]
-
E to add a constraint or a new column marked as
PRIMARY KEY is not supported. You can get pretty much the same
effect by doing:
CREATE UNIQUE INDEX custsales_i ON custsales(id, type);
Dan.
-
To unsubscribe, send email
sqlite3_bind_int. I couldn't figure out what is the
problem and how to fit it. It used to work last week.
Could you please help.
It's likely that the statement handle is invalid or has already
been passed to sqlite3_finali
One of the example tokenizers uses the ICU library for localization.
See the same document for details. It is built if the
SQLITE_ENABLE_ICU macro is defined when fts3 is compiled.
Regards,
Dan.
-
To unsubscribe, send email to [EMAIL PROTECTED]
-
ot;sqlite-amalgamation-3_5_4.zip" produced the same errors
as sqlite-source-.
Hi,
Can you try the experimental configure system in this package
and let us know if it works:
http://www.sqlite.org/sqlite-3.5.4-autoconf.tar.gz
Thanks,
Dan.
I have compiled, installed, linked, and used i
then
see if you can get a stack trace for another thread? Thanks.
Dan.
Ken <[EMAIL PROTECTED]> wrote: Core/backtrace info:
(gdb) core-file core
Core was generated by `/home/ixion/LIB/sqlite3test/.libs/
testfixture ../sqliteSrc/sqlite-3.5.5/test/qu'.
Program terminated with signal
On Feb 6, 2008, at 1:00 PM, Pavel Dedik wrote:
> Hi all,
>
> how can I escape the '%' in where clause?
> I suggested the query may look like this:
>
> *SELECT * from tbl where field LIKE '%\%text';
Try:
select * from a w
he incorrect HTML is equivalent to the HTML
suggested by the validator. Fair enough I guess, both versions
are unambiguous and the suggested version is better.
Dan.
___
sqlite-users mailing list
sqlite-users@sqlite.org
http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
an atomic
operation (either all the changes will succeed or none
of them). You can't do that with two separate database
handles.
Dan.
> I was initially thinking that I could attach every database I had
> with the
> same schema and execute queries without specifying database names.
On Feb 8, 2008, at 11:31 AM, Rob Sciuk wrote:
>
> How do I sign up for the digest rather than the regular feed??
>
>
> http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
Use the webpage linked above.
Dan.
___
sqlite-use
e resource that really does require a destructor (a
connection handle to some other database etc.), do not open it in
xBestIndex(). Open it in xFilter() and close it in xClose(). Each
xFilter() call should be matched by exactly one xClose().
Regards,
Dan.
_
> }
> }
>
>
> So my questing is, does sqlite allow a prepare to be started while
> a prior prepared statement is still open?
Hi,
This is supposed to be Ok. If you can supply an example program
to demonstrate the crash that would
t databases, even those with
the same schema. If the table/index locations are different
but the schema cookie is the same, you are risking database
corruption.
Also, prepared statement structures contain some pointers
back to the database handle that was used to create them.
Not sure what the implicati
base? Or send it to
me if you don't want it made public. Also, the exact sqlite version
might be helpful.
Alternatively, can you compile with debugging symbols enabled and
post a stack trace?
Thanks,
Dan.
On Mar 3, 2008, at 4:33 AM, P Kishor wrote:
> I have a fairly simple db with fts3
but some others show it being called from other
parts of SQLite that make no sense to me. This one is particularly
odd:
http://crash-stats.mozilla.com/report/index/8b54f1c5-e8aa-11dc-
b466-001a4bd43e5c
Stack overflow possibly? Will keep thinking this.
Regards,
Dan.
___
On Mar 3, 2008, at 12:03 PM, P Kishor wrote:
> Dan,
>
>
> On 3/2/08, Dan <[EMAIL PROTECTED]> wrote:
>>
>> Hi,
>>
>> I tried this script here with the latest CVS version and
>> it didn't crash:
>>
>>CREATE TABLE page
ld understanding that sqlite3_get_table (previously
> sqlite_get_table) will malloc for *errmsg and I'll need to do the
> free(). Has anything changed? I'm just recompiling old codes with
> the new lib.
Technically you're supposed to use sqlite3_free().
On Mar 10, 2008, at 6:44 PM, thomas Armstrong wrote:
> Hi.
>
> I've just installed 'sqlite' (version 2.8.16) and 'sqlite-devel'
> (version 2.8.16) from RPMs on my Linux Fedora Core, but I get this
> error message when installing a software:
Looks
3_result_text() to return the result, does your
result string include a nul-terminator character? If so, that byte
should not be included in the "number of bytes" parameter passed
to result_text(). i.e. if you were doing:
sqlite3_result_text(pContext, "abc", 4, ...)
you might get
On Mar 11, 2008, at 3:02 PM, [EMAIL PROTECTED] wrote:
>
>
> OK Dan, you have the solution. The count was including the
> terminating NULL char. Making it not include the NULL char fixed
> the problem.
>
>Another question: For an empty result, should I return 0 or
lite3_mprintf("INSERT OR IGNORE INTO ..."...
^
Maybe that's got something to do with it. Knock out the OR IGNORE
and see what happens.
Dan.
___
sqlite-users mailing list
sqlite-users@sqlite.org
http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
ext_%s WHERE
> contents
> MATCH '? '",sTempTableRoot.c_str());
Don't put quotes around the ? character. ('? ') is a literal string. You
probably want (? || ' ') - an SQL variable with a single space appended
to it.
Dan.
>
> nError = sqlite3_prep
> How does SQLite handle sorting with a column that contains values
> (including zero) and NULs? Are the NULs converted to zero for the
> purposes of sorting?
A null value is considered less than all other values when sorting.
http://www.sqlite.org/datatype3.html#compari
; anyone have
> an example using this function (sqlite3_create_function) to override
> regexp()?
There is one in the ext/icu/icu.c file of the source distribution.
http://www.sqlite.org/cvstrac/fileview?f=sqlite/ext/icu/icu.c&v=1.7
Dan.
_
sodes fe2
> Why does this put SQLite into an endless loop?
Why do you figure it is an infinite loop? From the looks of the
query it is probably just returning a very large number of rows.
Dan.
> (I am not using the shell tool, just preparing the statement as is
> and
for page tracking?
Only the dirty-page tracking. The limits on the number of
attached databases and tables in a join are the same as
they always were.
Dan.
___
sqlite-users mailing list
sqlite-users@sqlite.org
http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
nds
the ICU library to SQLite using the custom collation sequence interface
(and others). With this extension, SQLite uses the upper/lower case
tables that are part of unicode.
See ext/icu/README in the source distro for details.
Dan.
> Keith.
>
>>> Sort order is highly depende
static sqlite3_io_methods njMethods = {
> 1,
> njOpen, ... } ;
>
> }
>
>
> How do I get the address of sqlites default I/O methods?
sqlite3_vfs *pDefaultVfs = sqlite_vfs_find(0);
Dan.
>
> Thanks for any sug
On Apr 8, 2008, at 2:16 AM, Ken wrote:
> Dan,
>
> Doesn't that just return the VFS part ??? I'm looking for the
> sqlite3_io_methods part for the default vfs ( or is this really
> the sqlite3_file )?
Right. Sorry, I didn't read carefully. :)
I don't rea
On Apr 12, 2008, at 12:36 AM, dark0s dark0s wrote:
> Excuse me for my stupid topic, but I am crazying to find error in
> program below.
>
> My output is:
>
> bash-3.1# gcc -shared labsinf.c -o inf.so
> labsinf.c:61: error: expected ';', ',' or ')' before '*' token
>
> The errror is for sqlit3_e
emory available to an application
is all of the RAM + all of the swap space. Best performance
probably comes by using up all of the RAM and never using the swap.
Realistically, you should probably just set a large cache_size as
in option 1. Does SQLite really run faster with 1GB available than
it would with 100MB?
Dan.
___
sqlite-users mailing list
sqlite-users@sqlite.org
http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
>
> I was not so much concerned between zip vs .tar.gz format, but the
> content of the amalgamation.
"sqlite-amalgamation-3.5.8.tar.gz" which is the file that allows
you do "./configure && make install" to build from the amalgamation
is up now.
Dan.
__
asier to read.
> In any case, would I be better off with a compound index if I modify
> my query to
>
> select wkt_geometry from sg_rivers where xmin >= ? and ymin >= ? and
> xmax <= ? and ymax <= ?;
No. The B-Tree structures that SQ
dll
> unloading and I traced it back to one of sqlite functions. I modified
> (hacked) it for myself to be able to free the slot on request, but it
> would be nice if future version addressed this problem.
Upgrading will probably clear this problem. As of 3.5, SQLite no
long
ads don't operate on the same db at the same
> time?
As of 3.5, sqlite connections are threadsafe by default. With
earlier versions, this trick will not work.
Dan.
___
sqlite-users mailing list
sqlite-users@sqlite.org
http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
ely into its own attached db?
>
>
> in other wordstwo bulk insert operations going on simultaneously
> on the same connection but each insert operation going into a
> different attached in-memory db.
Probably not. Each sqlite3* handle has a single mutex that it uses
to serializ
We don't have any steps to reproduce, but we have a bug tracking
> the issue:
> https://bugzilla.mozilla.org/show_bug.cgi?id=432148
>
> More details can be found in the bug.
Hi Shawn,
Do these databases use auto-vacuum? Incremental vacuum? Fts3?
Thanks,
Dan.
>
&
, but the theoretical
worst performance is terrible.
If you are using automatically generated rowids for your table (you
are unless you have declared a column as INTEGER PRIMARY KEY) and
the indexed timestamps are always increasing (I guess they are,
right?) then
.html
> (what would we do without Google !?!)
>
> Is it possible to include this (or a better equivalent) to the next
> version? I hate such hacks where I don't really know what I am doing.
You might have some luck if you grab cvs as of last night. Some
changes were m
ring is
> still visible in the
> database file, the file instead grows. What is the rule for reusing
> a deleted block (if it
> is so)?
I guess it depends on the size of the new records inserted how
space is allocated for them. Assuming they are sm
y. One of the most annoying this about
> SQLite is that version 3.x can't open version 2.x databases (which
> unfortunately are still out there).
I think the short answer is that with version 4:
* Descending indexes are properly supported, and
* Integer values 0 and 1 consume a s
On May 17, 2008, at 2:51 AM, Shawn Wilsher wrote:
> And a fun follow-up question. Will sqlite3_transfer_bindings transfer
> bindings across connection objects if the two statements are for two
> different connections to the same database?
No. It will return SQLITE_MISUSE.
Dan.
&
1 - 100 of 1918 matches
Mail list logo