--- [EMAIL PROTECTED] wrote:
> Joe Wilson <[EMAIL PROTECTED]> wrote:
> > --- [EMAIL PROTECTED] wrote:
> > > It may be more difficult to implement this in a backwards-compatible
> > > way such that older versions of SQLite can rollback a journal created
> >
Unfortunately, the sqlite3 commandline shell is not part of the test suite
despite its widespread use as an administrative tool for sqlite databases.
http://marc.info/?l=sqlite-users=117253099812346=2
But you know this already - you're the same guy as this previous post. :-)
--- Travis
> What is worse is that VACUUM didn't really help that much. It takes
> forever, and it doesn't really "fix" the fragmentation either.
That used to be the case, but VACUUM is vastly improved in the latest
version of SQLite.
--- Mike Johnston <[EMAIL PROTECTED]> wrote:
> Are there any significant performance benefits by limiting the number of
> columns in a single
> table select statement? Does joining (<5 tables) make a significant
> difference to that answer?
If you need the columns, you don't have much choice
--- [EMAIL PROTECTED] wrote:
> Consider this query:
>
>SELECT a, b FROM t1 UNION SELECT b, a FROM t1 ORDER BY a,b;
>
> Is the query above equalent to:
>
> (1) SELECT a, b FROM t1 UNION SELECT b, a FROM t1 ORDER BY 1,2;
>
> Or is it the same as:
>
> (2) SELECT a, b FROM t1 UNION
--- "Andrey A. Lapin" <[EMAIL PROTECTED]> wrote:
> I develop database in which information on the traffic (headings of packets
> MAC, TCP/IP) is stored.
> Processing of queries occurs slowly.
> For example, quantity of records in a database of 17 million lines and its
> size 3 Gb.
> How to
--- DragonK <[EMAIL PROTECTED]> wrote:
> I'm having the following problem: a sqlite database file is on an NTFS
> filesystem, in a directory with no permissions to create new files, but only
> to modify the original database. By using filemon i've noticed some access
> denied errors when sqlite
--- [EMAIL PROTECTED] wrote:
> I'm using cygwin under windows XP.
> gcc version 3.4.4
>
> I unzipped the sqlite-3.3.16.tar.gz to the directory sqlite-3.3.16.
>
> Executed the following:
>
> cd sqlite-3.3.16
> mkdir build
> cd build
> ./configure
> make
>
> The resulting sqlite3.exe is 4 times
--- Miha Vrhovnik <[EMAIL PROTECTED]> wrote:
> > strip sqlite3.exe
> is not ok at least for DLL, because it strips everything even reallocation
> info.
> strip --strip-unneeded sqlite3.dll
The GNU toolchain's strip command corrupts the reallocation information for
MinGW DLLs, so I would not
> I've used SQLite-3.3.4 for quite a while, mostly the Tcl binding.
> A few days ago I installed 3.3.15 and the Tcl binding worked fine.
> Tonight I downloaded 3.3.16 and compiled it without any errors
> or warnings and then installed it. When I tried to execute sqlite3
> the following error
> I've used SQLite-3.3.4 for quite a while, mostly the Tcl binding.
> A few days ago I installed 3.3.15 and the Tcl binding worked fine.
> Tonight I downloaded 3.3.16 and compiled it without any errors
> or warnings and then installed it. When I tried to execute sqlite3
> the following error
--- [EMAIL PROTECTED] wrote:
> Ulrich =?iso-8859-1?q?Sch=F6bel?= <[EMAIL PROTECTED]> wrote:
> > On Saturday 21 April 2007 15:40, Joe Wilson wrote:
> > >
> > > Apply this patch to fix this problem.
> > >
> >
> > Thanks for the patch, but t
--- Joe Wilson <[EMAIL PROTECTED]> wrote:
> In general, sqlite3.so is backwards compatible, but not necessarily
> forward compatible. So you can upgrade an older binary to use a more
> recent 3.x sqlite3.so release, but not the other way around.
I meant to to say:
In gene
I noticed that sqlite tar.gz downloads prior to
http://sqlite.org/sqlite-3.3.10.tar.gz
are missing. So historical releases are not available except by CVS.
Is this intentional?
__
Do You Yahoo!?
Tired of spam? Yahoo! Mail has the best spam
--- Ulrich Schöbel <[EMAIL PROTECTED]> wrote:
> Thanks for the patch, but there is still something wrong in the
> linking stage.
>
> I have 3 instances of SQLite on my system:
> The first is 3.2.1, installed with the system and not used, at least
> not by me, library in /usr/lib.
> The second is
--- [EMAIL PROTECTED] wrote:
> Joe Wilson <[EMAIL PROTECTED]> wrote:
> > I noticed that sqlite tar.gz downloads prior to
> >
> > http://sqlite.org/sqlite-3.3.10.tar.gz
> >
> > are missing. So historical releases are not available except by CVS.
> &
This libtool --tag problem does not seem to be unique to sqlite.
http://lists.freebsd.org/pipermail/freebsd-questions/2005-November/104076.html
I think it's a question for the freebsd port maintainers.
--- snowcrash <[EMAIL PROTECTED]> wrote:
> hi,
>
> building v3.3.16 from src on osx works
The test script itself has the test.db open, and as result Windows cannot
delete an open file. It worked on UNIX because you can delete anything at
any time whether it is open, running, locked or whatever.
If you apply this patch, exclusive2.test will run to completion on cygwin
without error.
a better patch...
Index: test/exclusive2.test
===
RCS file: /sqlite/sqlite/test/exclusive2.test,v
retrieving revision 1.4
diff -u -3 -p -r1.4 exclusive2.test
--- test/exclusive2.test16 Apr 2007 15:02:20 - 1.4
+++
What's an ads1.2?
--- [EMAIL PROTECTED] wrote:
> Dose anybody compile the sqlite3 in the ads1.2 environment?
__
Do You Yahoo!?
Tired of spam? Yahoo! Mail has the best spam protection around
http://mail.yahoo.com
Try compiling with sqlite3.c from http://sqlite.org/sqlite-source-3_3_17.zip
It is already pre-generated and does not require generating the parser from
the .y file.
See: http://www.sqlite.org/cvstrac/wiki?p=TheAmalgamation
--- [EMAIL PROTECTED] wrote:
> arm developer suite v1.2
> support
> When I try to use the header I get errors
>
> [C++ Error] sqlite3.h(1778): E2232 Constant member
> 'sqlite3_index_info::nConstraint' in class without constructors
It appears it is trying to compile the sqlite header file as if it were C++.
Lemme guess - you're using the almalgomated
I wrote too quickly - sqlite3.h correctly uses __cplusplus for extern "C".
But the almalgomated sqlite3.c cannot be compiled from a C++ compiler
for the reasons described below.
--- Joe Wilson <[EMAIL PROTECTED]> wrote:
> > When I try to use the header I get error
--- Jonathan Kahn <[EMAIL PROTECTED]> wrote:
> I really appreciate your response. What do you suggest I do? Is there
> something else I need to include aside from sqlite3.lib? I am willing to
> try anything.
I only use GNU C++, so I can't help you with .lib files.
I'd suggest to compile
> create table filemap(id integer primary key,
>uid integer, gid integer, mtime integer,
>vol integer,
>path varchar(1024));
>
> It has no indices built yet.
>
> I'm adding quite a lot of records to it using a perl script which
> generates SQL like this:
>
> begin;
> insert into
--- Joe Wilson <[EMAIL PROTECTED]> wrote:
> > create table filemap(id integer primary key,
> >uid integer, gid integer, mtime integer,
> >vol integer,
> >path varchar(1024));
> >
> > It has no indices built yet.
> >
> > I'm a
--- Tim Bradshaw <[EMAIL PROTECTED]> wrote:
> > I can't reproduce your problem. I can insert 16M records into your
> > table
> > schema in 25 minutes on a 5 year old Windows machine. The sqlite3
> > process
> > had peak RAM usage of less than 20M.
>
> Rats, I suspect it must be some
--- Ron Stevens <[EMAIL PROTECTED]> wrote:
> Is it possible to tell SQLite to limit the size that a database may
> grow to? It would be useful for storage constrained applications.
This is a tricky problem.
What would you have the database do if an insert failed upon reaching
the limit?
What
I wrote too soon:
http://www.sqlite.org/cvstrac/chngview?cn=3941
+ /*
+ ** Maximum number of pages in one database file.
+ */
+ #ifndef SQLITE_MAX_PAGE_COUNT
+ # define SQLITE_MAX_PAGE_COUNT 1073741823
+ #endif
--- Joe Wilson <[EMAIL PROTECTED]> wrote:
> --- Ron Stevens <[EMA
--- [EMAIL PROTECTED] wrote:
> Joe Wilson <[EMAIL PROTECTED]> wrote:
> > I wrote too soon:
> >
> > http://www.sqlite.org/cvstrac/chngview?cn=3941
> >
> > + /*
> > + ** Maximum number of pages in one database file.
> > + */
> > + #ifndef S
Does anyone know if this bug was the result of a recent btree optimization,
or was it a longstanding issue?
Ready
for the edge of your seat?
Check out tonight's top picks on Yahoo! TV.
> The DB schema is:
> CREATE TABLE xxx(a TEXT, b TEXT, c TEXT);
> CREATE INDEX idx on xxx(b);
>
> The command
> explain select rowid, a, b, c from xxx where b in('qwerty') order by rowid;
> shows 58 opcodes.
>
> The command
> explain select rowid, a, b bbb, c from xxx where bbb in('qwerty')
--- Joe Wilson <[EMAIL PROTECTED]> wrote:
> > The DB schema is:
> > CREATE TABLE xxx(a TEXT, b TEXT, c TEXT);
> > CREATE INDEX idx on xxx(b);
> >
> > The command
> > explain select rowid, a, b, c from xxx where b in('qwerty') order by rowid;
Is the sole purpose of NameContext.nRef to see whether you're dealing
with a correlated subquery?
We won't tell. Get more on shows you hate to love
(and love to hate): Yahoo! TV's Guilty Pleasures list.
--- Juri Wichanow <[EMAIL PROTECTED]> wrote:
> > In my database of 30 000 000 records.
> > At " Pragma cache_size = 100 " indexation lasts > 12 hours,
> > at " Pragma cache_size = 2000 " - 45 minutes.
It would be interesting if you ran a profiler such as gprof during
each run to see
--- "Sergey M. Brytsko" <[EMAIL PROTECTED]> wrote:
> Hi All!
>
> The DB schema is:
> CREATE TABLE xxx(a TEXT, b INTEGER, c TEXT);
> CREATE INDEX idx on xxx(b);
>
> explain query plan select rowid, a, b, c from xxx where b > 1 order by
> rowid;
> 0|0|TABLE xxx USING PRIMARY KEY ORDER BY
>
No error when run with most recent SQLite.
replace() may not have existed in 3.3.7.
--- Jim Dodgen <[EMAIL PROTECTED]> wrote:
> I get an error in version 3.3.7 when using the replace function as defined
> here:
>
> http://sqlite.org/lang_expr.html
>
> "replace(X,Y,Z) Return a string
This updated patch greatly improves query times against compound
SELECT statements (i.e., UNIONs) when the parent SELECT has a WHERE
clause.
For example, this query:
SELECT * FROM (
SELECT a,b FROM t1
UNION ALL
SELECT x,y FROM t2
) WHERE a>b;
will be transformed into the more
Improved patch against latest CVS with more comments and new test
case attached. No regressions with make test.
> This updated patch greatly improves query times against compound
> SELECT statements (i.e., UNIONs) when the parent SELECT has a WHERE
> clause.
In a previous post drh mentioned:
You need an R-Tree index to do something like this. The
public-domain version of SQLite only supports B-Tree indices.
(http://www.mail-archive.com/sqlite-users%40sqlite.org/msg24643.html)
Does this imply that there exists a commercial version of SQLite
If you're using Linux, try:
echo "create table t1(a);" |strace ./sqlite3 my.db 2>&1 |less
and examine the output. See where it differs from the successful
tmpfs run.
Newer versions of sqlite3 may have better IO error messages.
> I'm running the command line tool to create sqlite3 db.
> My
Nothing you've mentioned is out of the ordinary. I would expect the
same behavior on both platforms.
Can you post the complete schema, and the exact query that exhibits the
problem? (And perhaps a couple of insert statements into the objects table).
Without this I don't think anyone can recreate
Just for kicks, what happens on both platforms when you issue:
select genre, length(genre), hex(genre), filename
from objects where media_type=1;
as well as:
select count(*) from objects where genre LIKE '%POP%';
> I have a bizarre problem. Here is an example of something I tried in
>
What is the OS you're using and what kind of media is it?
Hard drive or USB key or ???
--- Shilpa Sheoran <[EMAIL PROTECTED]> wrote:
> It seems that rc = fsync(fd); is failing in function
> static int full_fsync(int fd, int fullSync, int dataOnly) in file os_unix.c
> {
> #else /* if
--- "Tian-Jian \"Barabbas\" [EMAIL PROTECTED]" <[EMAIL PROTECTED]> wrote:
> Hi Millan,
>
> I encountered the same problem, the record is on
>
> http://b6s.blogspot.com/2007/04/previous-version-ghost-on-vista-can-be.html
>
> and the issue was sent to
>
>
--- Nigel Metheringham <[EMAIL PROTECTED]> wrote:
> I recently experimented with putting binary data into SQLite table
> rows - I declared the column holding the binary data as a BLOB (not
> that it makes a ton of difference for SQLite).
>
> This worked very well (using perl
> I've a "black-box" (Linux) program - so no source available - using a
> sqlite database. Is there any way I can monitor/log the queries made to
> the database?
If it's dynamically linked, just replace libsqlite3.so with your own.
If it is statically linked and has -g symbols, use gdb.
If it's
--- Lloyd <[EMAIL PROTECTED]> wrote:
> On Wed, 2007-05-23 at 07:21 -0700, Joe Wilson wrote:
> > > I've a "black-box" (Linux) program - so no source available - using
> > a
> > > sqlite database. Is there any way I can monitor/log the queries made
--- Nigel Metheringham <[EMAIL PROTECTED]> wrote:
> I'll see if I can reproduce the result with standard INSERT statements
Try this:
sqlite3 orig.db vacuum -- might be necessary to preserve row order
sqlite3 orig.db ".dump BrokenTable" | tee orig.sql | sqlite3 new.db
sqlite3 new.db ".dump
--- Brett Keating <[EMAIL PROTECTED]> wrote:
> Msica independiente|text|19|0056_People Get Ready1_test1.wma
> POP|text|3|0057_The Mighty Ship1_test1.wma
> POP|text|3|0058_The Mighty Quinn1_test1.wma
>
> Anyway, it turns out the problem was caused by creating an index on the
> genre field. If I
--- Brett Keating <[EMAIL PROTECTED]> wrote:
> Actually this wasn't the issue after all... Indices have nothing to do
> with it.
>
> The genre was being inserted from two different sources. It is a UTF-16
> string, and in one case it was being inserted with a null terminator,
> and in another
--- Alexander Smondyrev <[EMAIL PROTECTED]> wrote:
> I am trying to use loadable extensions in Sqlite and I've run into the
> following 2 problems:
>
> 1) I've downloaded src for 3.3.17 Sqlite and build it, but the '.load'
> option does not seem to appear when I run the shell. I've used the
On Windows, timing sending the results to a file:
sqlite3 your.db "SELECT * FROM trend_data" > foo.txt
On a 2.0 GHZ P4 I get 50,000 rows/second, which is not bad considering
it's a 5 year old machine and sqlite3 was not compiled with much
optimization.
Any new machine should be twice as
--- Leif Jensen <[EMAIL PROTECTED]> wrote:
>In a larger project we are using PostgreSQL database and Embedded SQL
> in C (using ECPG) in a server daemon. We would like to be able to have this
> to work with
> SQLite for a stand-alone application. The Embedded SQL in C standard is
> as used
--- Leif Jensen <[EMAIL PROTECTED]> wrote:
> I thought about the "fake" library myself
> even though our project is more like 20 - 30k lines, but I'm not sure
> how bad it would be.
May not be worth the hassle given the size of your project.
Since you're using C, consider rewriting your code
> I am working at porting sqlite ( ver 3.3.8 ) on an embedded device with
> extremely low main memory.
>
> I tried running select queries on the tables( with about 2k records each
> having about 5 strings) and they do well within 20kB of runtime heap
> usage.
>
> But, when I try new insertions,
--- Joe Wilson <[EMAIL PROTECTED]> wrote:
> > I am working at porting sqlite ( ver 3.3.8 ) on an embedded device with
> > extremely low main memory.
> >
> > I tried running select queries on the tables( with about 2k records each
> > having about 5 strings) and
ndows, in my opinion.
--- Kalyani Tummala <[EMAIL PROTECTED]> wrote:
> My temp_store is SDRAM. Thanks for your suggestion of using COMMIT. I
> have not used it. Any other pointers?
>
> Best Regards
> Kalyani
>
>
> -Original Message-
> From: Joe Wilson [ma
I think I know what's going on.
When you insert new rows in the presence of indexes then sqlite must
touch a lot of pages in each trascation to satisfy the rebuilding of
the index(es). These pages are built up in the transaction log which is
stored in temp_store, which happens to be memory in
> I'd like to figure out why the example
> with half function which was provided by sqlite team can't be loaded as a
> shared library.
Yeah, it's poorly documented. I'll just put this in the wiki:
http://www.sqlite.org/cvstrac/wiki?p=LoadableExtensions=1180475067=1
How To Build a Loadable
It may not be possible to get peak heap usage down to 30K, but here's
some random ideas:
I imagine you've already tried defining SQLITE_OMIT_* for the features
that you don't need.
Verify that your embedded OS has a space-efficient malloc implementation.
Try to find a realtime graphical heap
You could save a few bytes in some sqlite internal structs if you'd
use C bitfields for boolean flags:
For example:
struct MemPage {
u8 isInit; /* True if previously initialized. MUST BE FIRST! */
u8 idxShift; /* True if Cell indices have changed */
u8 nOverflow;
struct MemPage2 {
u8 nOverflow;/* Number of overflow cell bodies in aCell[] */
u8 childPtrSize; /* 0 if leaf==1. 4 if leaf==0 */
u8 hdrOffset;/* 100 for page 1. 0 otherwise */
u8 isInit:1; /* True if previously initialized. MUST BE FIRST! */
Okay, maybe not
MemPage bitfield patch below.
sizeof(MemPage) on Linux:
original: 84
patched: 76
Patched "make test" runs without regressions on Linux and Windows.
Timings for "make test" (elapsed):
original: 1:20.74
patched: 1:20.22
Size of sqlite3.o when compiled from almalogmation with all
--- Dennis Cote <[EMAIL PROTECTED]> wrote:
> There are also some issues with regard to the ordering and layout
> of bitifleds in cross platform applications. I suspect that is the
> reason they aren't used.
If an external interface changed, sure. But these internal structs
change constantly
--- Dennis Cote <[EMAIL PROTECTED]> wrote:
> Joe Wilson wrote:
> >
> > If an external interface changed, sure. But these internal structs
> > change constantly from (minor) release to release.
> >
> > The struct in question is used solely by btree.c, so the
> > If the MemPage are malloced individually (instead of being put in arrays),
> > then they are 16
> byte
> > aligned on most platforms, making the allocated block effectively the same
> > size (well, that
> > depends on how many bytes are used by malloc before the user block in
> > memory).
>
Be aware of a Windows OS bug that prevents correct conversion of epoch
integers to local date/time due to the recent US DST change:
http://www.sqlite.org/cvstrac/tktview?tn=2322
Assuming you've applied the Windows OS DST patch, epoch-converted
times can be off by an hour for pre-2007 dates in
--- John Stanton <[EMAIL PROTECTED]> wrote:
> Sqlite lets you put in anything as the declared type. "DEAD PARROT",
> "MONGOOSE", "GODZILLA" or "DECIMAL(6,1)" are all acceptable declared
> types. Sqlite makes the underlying type TEXT if it is not obviously
> numeric.
The default affinity type
--- Mitchell Vincent <[EMAIL PROTECTED]> wrote:
> I guess this isn't possible after all?
Get the SQLite ODBC driver source code and alter it to do whatever
you like when the type "timestamp" column comes up.
> On 5/31/07, Mitchell Vincent <[EMAIL PROTECTED]> wrote:
> > I have a set of databases
--- Dan Kennedy <[EMAIL PROTECTED]> wrote:
> On Fri, 2007-06-01 at 10:51 +0100, Mark Gilbert wrote:
> > Folks.
> >
> > My app just crashed in the field randomly after some time running fine.
> >
> > Thread 12 Crashed:
> > 0 libsqlite3.0.dylib 0x9406e587 sqlite3pager_get + 390
> >
--- John Stanton <[EMAIL PROTECTED]> wrote:
> Sqlite does have a date format, it is physically a 64 bit floating point
> number. There are functions to transform in and out of that format to
> present dates as required by the user. The Sqlite date format uses a
> magib epoch which matches all
--- [EMAIL PROTECTED] wrote:
> I'v read in change log that some stack allocted memory were moved to the
> heap, but I think that
> there is still to much allocated memory on the stack.
> After creating a table with 2000 columns, jdbc driver created a query that
> run out of stack.
> Default
Such a statement would never be issued on a low memory device.
This is an exceptional case involving a select with 2000
unions - I would not worry about it.
--- [EMAIL PROTECTED] wrote:
> This is very worrying since it means that the statement cannot be compiled on
> a
> low memory device.
> I
--- [EMAIL PROTECTED] wrote:
> Joe Wilson napisa³(a):
> > Please respond to the mailing list in the future.
>
> Sorry. Different client. I didn't notice the adress.
>
> > At least there's a known workaround, so no problem.
>
> Workaround is not a solution.
Incr
--- [EMAIL PROTECTED] wrote:
> My application's doesn't create any databases itself. It allows users to
> store any data. And users need to be able to store any number of columns in
> 1 table (the most I'v heard about is about 1, but I wouldn't be
> surprised if they had more). Trust me,
> A Linux 2.6/x86_64 system reports a "disk I/O error" (SQLITE_IOERR)
> while generating a specific report from a SQLite database (SQLite
> 3.3.6). The database and temporary files are accessed through an NFS
> mount. After running the program again with SQLite tracing enabled
> (plus a bit more
.dump
> when using a blob column in the command line tool 'sqlite3' im getting
> garbled output when selecting from a table that contains a blob column.
>
> Is there a way to get the blob column output in escaped format, like its
> used in the insert statement?
CREATE TABLE t(b blob);
INSERT INTO "t" VALUES(X'ABCD');
select quote(b) from t;
X'ABCD'
> Joe Wilson wrote:
> > .dump
>
> yes, but dump gives me all rows for the table and not those that are the
> result of a query (which might have a WHERE clause).
>
> Is
You're crashing in free(), which means your heap is corrupted.
The cause of the corruption could be from anywhere - and not
necessarily sqlite. It might be the victim of a previously corrupted
heap. Try running your program through a memory checker like valgrind
to see what it turns up.
Also
--- Guido Ostkamp <[EMAIL PROTECTED]> wrote:
> On Sat, 9 Jun 2007, Joe Wilson wrote:
> > CREATE TABLE t(b blob);
> > INSERT INTO "t" VALUES(X'ABCD');
> > select quote(b) from t;
> > X'ABCD'
>
> Thanks, Joe. I am just wondering why this 'quote' is n
> $sql = "SELECT count(*) FROM feedback";
> $result = $handle->query($sql);
> echo "rows = " . $result->rowCount() . "\n";
I've never used PHP, but just for kicks, to eliminate
the database from the equation, try:
$sql = "SELECT 123 AS abc;";
and see what happens.
--- Olaf Schmidt <[EMAIL PROTECTED]> wrote:
> On what OS have you tested?
> If on windows, was it a GCC-compile or a MS-VC-compile?
Running the GCC cross-compiled sqlite3.exe from
http://www.sqlite.org/sqlite-3_3_17.zip on Windows
(well, wine on Linux):
SQLite version 3.3.17
Enter ".help"
--- [EMAIL PROTECTED] wrote:
> Going into this statement on both Linux and Windows,
> the exact same values are in realvalue and in rounder:
>
>realvalue: 0xf000 0.94995559107901499373838
>rounder:0x3faa 0.0500030
>
This is a question for your OS vendor.
--- "Rachmel, Nir (Nir)" <[EMAIL PROTECTED]> wrote:
> I have run into a problem - using the same environment variables as I
> did in the old version ( I am upgrading from 3.2.8 to the newest version
> of 3.3.17), I am unable to compile succesfully.
> Here is
> TabA.ID1
> TabA.ID2
> TabA.field1
>
> TabB.ID1
> TabB.ID2
> TabB.field2
>
> TabC.ID1
> TabC.field3
>
> ViewBC:
> SELECT * FROM TabB INNER JOIN TabC On TabB.ID1 = TabC.ID1
>
> This is slow:
> SELECT field1, field2, field3 from TabA LEFT OUTER JOIN ViewBC ON TabA.ID1 =
> ViewBC.ID1 AND TabA.ID2
Large bulk inserts with more than one index (implicit or explicit)
is not SQLite's strong suit.
If you search the mailing list archives you'll find a few suggestions:
- "BEGIN EXCLUSIVE" (or is it "BEGIN IMMEDIATE"?) on the
database file and then copy the file over - fastest way
or
-
> I tried to set the cache size to 0 (after sqlite3_open0, and then query
> for pragma cache_size which returns 2000 (default cache size). Why its
> not returning 10 (according to Weiyang Wang)?
It does report 0, even though internally it is using a value of 10.
SQLite version 3.3.17
Enter
> If the performance problem is with the seconday index, is there a way to
> "pause" indexing before
> a large bulk insert and then "resume" it later without rebuilding the entire
> index (to avoid
> doing: drop index + inserts + create index)?
No
--- Omar Eljumaily <[EMAIL PROTECTED]> wrote:
> Is it possible to get the name of the function that the callback was
> called because of?
If you specified a userData argument in sqlite3_create_function()
you can retrieve it in the functon itself with
void
--- [EMAIL PROTECTED] wrote:
> I have a user-defined function named DECRYPT, which decrypts column data
> encrypted by my other UDF named ENCRYPT.
>
> The UDF callback function (which does the decrypting) calls
> sqlite3_result_blob
> after decrypting the data. Sqlite does return the data
If both field types were BLOB_TEXT, then that should be provided. Only
> in cases
> where there were multiple columns and the data types were different
> would a NULL
> be returned.
>
> This would certainly be an improvement for Sqlite, would it not?
>
> cheers
> -brett
--- [EMAIL PROTECTED] wrote:
> Joe Wilson <[EMAIL PROTECTED]> wrote:
> >
> > As a possible extension one could see sqlite3_create_function taking an
> > optional argument with a hint as to its return type that sqlite may use
> > for sqlite3_column_decltype. But
--- Scott Hess <[EMAIL PROTECTED]> wrote:
> On 6/14/07, Joe Wilson <[EMAIL PROTECTED]> wrote:
> > You can't infer a function's return type from its arguments.
> > Take the hypothetical function FOO(x). If I pass it a number, it will
> > return the number s
Is there an SQLite 3.x equivalent document for this?
SQLite 2.X Database File Format
http://sqlite.org/fileformat.html
If not, is this 2.x document worth reading as a background to
the general structure of the sqlite 3.x file and page format?
Or has it changed so much that it's not useful?
> select tableA.path, tableA.value from tableA,tableB where
> tableA.path=tableB.path and (tableA.value=tableB.value or
> (tableA.value IS NULL AND tableB.value IS NULL));
>
> It's possible that won't use an index, either, due to the OR, in which
> case you could try a union between a select with
--- [EMAIL PROTECTED] wrote:
> Gerry Snyder <[EMAIL PROTECTED]> wrote:
> > [EMAIL PROTECTED] wrote:
> > > SQLite version 3.4.0 is now available for download
> > >
> >
> > The tcl bindings for windows appear to be missing. Is this deliberate?
> >
>
> Build-script bug. Now fixed. Try
--- Sean Cunningham <[EMAIL PROTECTED]> wrote:
> I have very large datasets and have found that the built in union,
> intersect, and except operations do not seem to use indices
> (would be happy to be proven wrong here). As such, they
> are not very speedy with large large data sets.
A
Does every single process (however insignificant) that reads or writes
to that sqlite database file run on the same 16 processor machine?
> 16 Processor machine
> ~40Gb ram
> EMC storage
> Running a huge Oracle 10G database
> Running a 3rd party application that generates HUGE IO.
> Part of
--- "Sergey M. Brytsko" wrote:
> The problem is the index is NOT used for query:
> SELECT BBB FROM XXX WHERE BBB <> 100;
>
> but in case of query
> SELECT BBB FROM XXX WHERE BBB > 100;
> all is ok
...
> The indices are very important for me, how should I build these queries?
Say you have the
1 - 100 of 709 matches
Mail list logo
