I still have that problem.
can anyone help me in right direction?
___
sqlite-users mailing list
sqlite-users@sqlite.org
http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
John Machin wrote:
> I have developed a table which maps most latin-decorated Unicode
> characters into the non-decorated basic form.
This is a fascinating article by Sean Burke (a linguist) about
converting all Unicode characters into US-ASCII.
On 17/06/2009 11:52 AM, Dennis Cote wrote:
> Jens Páll Hafsteinsson wrote:
>> Closing and opening again did not speed up steps 1-4, it actually slowed
>> things down even more. The curve from the beginning is a bit similar to a
>> slightly flattened log curve. When I closed the database and
On Wed, Jun 17, 2009 at 11:52:45AM +1000, John Machin scratched on the wall:
> On 17/06/2009 6:17 AM, Hoover, Jeffrey wrote:
>
> > One other note, if you have a primary key whose value is continually
> > increasing your pk index can become imbalanced and therefore
> > inefficient.
>
> A B-tree
On 17/06/2009 1:19 AM, Christophe Leske wrote:
>>> So far , so good, but my client also expects ANY simplification of a
>>> character to be recognized:
>>> Cote d'azur for instance should return "Côte d'azur"
>>> or the Sao Paulo issue - how can a search for "Sao Paulo" return "Sào
>>> Paulo"
Jens Páll Hafsteinsson wrote:
> Closing and opening again did not speed up steps 1-4, it actually slowed
> things down even more. The curve from the beginning is a bit similar to a
> slightly flattened log curve. When I closed the database and started the test
> again, a similar curve appeared
On 17/06/2009 6:17 AM, Hoover, Jeffrey wrote:
> One other note, if you have a primary key whose value is continually
> increasing your pk index can become imbalanced and therefore
> inefficient.
A B-tree becomes imbalanced? How so?
http://www.sqlite.org/fileformat.html#btree_structures says:
On Jun 16, 2009, at 8:28 PM, Dave Toll wrote:
> Hello list
>
>
>
> I've noticed that if I call "PRAGMA cache_size=0", my database rows
> are
> still cached (very little disk access observed). If I call "PRAGMA
> cache_size=1" I can see that there is very little caching (disk access
>
Hello list
I've noticed that if I call "PRAGMA cache_size=0", my database rows are
still cached (very little disk access observed). If I call "PRAGMA
cache_size=1" I can see that there is very little caching (disk access
observed). Is there a minimum allowed cache_size setting? Does
On Jun 16, 2009, at 6:03 PM, Kalyani Phadke wrote:
>
> if I create dump using sqlite3.5.4 and recreate DB using that dump
> data
> using sqlite3.5.4,everything works fine. if I create dump using
> sqlite3.5.4 and recreate DB using that dump data using
> sqlite3.6.15,gets
> error.
>
> The
Hmmm, I know I've done this before. I'll have to dig deeper and find an example.
--
Joel LucsySeysan wrote:
When I used your code I get: No current row
If I use: rdr.Read();
before your code, I get the same exception as before.
___
sqlite-users
I'm aware of the limitations of the 92 Standard, as I too have suffered the
same fate. But one thing for sure, conforming to the standard certainly
makes migration between "Standard" implementations somewhat simpler.
I'm not speaking standardally but speaking logically. (i.e. Mr. Spock,
logical)
When I used your code I get: No current row
If I use: rdr.Read();
before your code, I get the same exception as before.
___
sqlite-users mailing list
sqlite-users@sqlite.org
http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
The rdr is:
SQLiteDataReader rdr;
___
sqlite-users mailing list
sqlite-users@sqlite.org
http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
On Tue, Jun 16, 2009 at 6:44 PM, Seysan wrote:
> String cmd = String.Format("SELECT filename,content FROM documents
> WHERE did={0}",did);
> contentCommand = sqlconn.CreateCommand();
> contentCommand.CommandText = cmd;
> rdr =
On Tue, Jun 16, 2009 at 6:44 PM, Seysan wrote:
> String cmd = String.Format("SELECT filename,content FROM documents
> WHERE did={0}",did);
> rdr = contentCommand.ExecuteReader();
What type is rdr? I believe it should be byte[]
--
Joel Lucsy
"The dinosaurs became
Hello,
I'm Using "System.Data.SQLite" in VS 2008 C#.
I'm using DataReader to read the data.
Everything works find except the BLOB.
I want to get that Blob and Write it to a File.
Blob data is mostly Picture and Music.
Here a bit of Code:
did-> is the document id.
String cmd =
Hi All,
I ran two queries:
1) select * from signature;
I didn't see the "SQL error: database disk image is malformed"
2) But if I ran the "select * from sig order by peerid;" then I have seen
the malformed
...
Hi All,
I have the table is defined as below:
CREATE TABLE `signature` (
`sig` char(50) NOT NULL,
`id' bigint(20) default '0',
But I have ran the folowing command:
.output mySelect
select * from signature;
then I didn't see NULL values in the mySelect file at all
But I ran
if I create dump using sqlite3.5.4 and recreate DB using that dump data
using sqlite3.5.4,everything works fine. if I create dump using
sqlite3.5.4 and recreate DB using that dump data using sqlite3.6.15,gets
error.
The text generated from .dump command of sqlite3 version 3.5.4.
CREATE TABLE
I'll settle for wishful thinking if that is what it takes to clone a table
schema with a Create...Select...statement.
Fred
-Original Message-
From: sqlite-users-boun...@sqlite.org
[mailto:sqlite-users-boun...@sqlite.org]on Behalf Of Igor Tandetnik
Sent: Tuesday, June 16, 2009 12:23 PM
This may be a red herring, but how do you generate the 'random' characters for
the test? Could entropy exhaustion affecting the results?
Just a thought...
*** Doug Fajardo
-Original Message-
From: sqlite-users-boun...@sqlite.org [mailto:sqlite-users-boun...@sqlite.org]
On Behalf Of P
I vacuum frequently, particular after large updates or deletes.
Two other potential optimizations of inserts (you may already be doing
this):
- use bulk insert
- encapsulate the inserts within a transaction
One other note, if you have a primary key whose value is continually
increasing your pk
VACUUM cleaned up the file in my current test, after 1200 iterations,
making it run at ~4.6seconds again, rather than ~5.1. It seemed to get
it almost back to the performance of a clean file.
Didn't know about the vacuum command--Cool. By the way, the vacuum
operation takes ~1.6 seconds for my
On Tue, Jun 16, 2009 at 2:51 PM, Jens Páll
Hafsteinsson wrote:
> Closing and opening again did not speed up steps 1-4, it actually slowed
> things down even more. The curve from the beginning is a bit similar to a
> slightly flattened log curve. When I closed the database and
On Tue, Jun 16, 2009 at 1:59 PM, Jens Páll
Hafsteinsson wrote:
> Ok, but you might have to run this a bit longer (or more iterations).
I ran the code a bit longer, and yes, I do notice an increase. Here
are my results after 610 iterations... I have provided the numbers for
10
D. Richard Hipp wrote:
> It was done to fix a bug.
>
> SQLite allows arbitrary text as the "datatype" of a column. So you
> could say (for example):
>
> CREATE TABLE t1(a "duh! ++ x hi, yall!(+123)" unique);
>
> And the datatype for t1.a would be "duh! ++ x hi, yall!(+123)". It
> used to
On Jun 16, 2009, at 3:10 PM, Kalyani Phadke wrote:
> But before entering the bug , I would like to find out how this
> happened. Whats the cause of it. Because I am not able to duplicate it
> all the time. IS there any way to find out what went wrong?
> Thanks,
>> I have to go manually and edit
Wes, I gather you are using version 3.6.14 in this test (?)
Please also try version 3.6.15 since I'm seeing drastically different behavior
using that version (starting similarly fast as .14 but quickly slowing down
after about 15 iterations and then leveling off after 22 iterations).
JP
But before entering the bug , I would like to find out how this
happened. Whats the cause of it. Because I am not able to duplicate it
all the time. IS there any way to find out what went wrong?
Thanks,
-Kalyani
-Original Message-
From: sqlite-users-boun...@sqlite.org
Thanks for the info Konrad. This is very similar to what I'm doing so that
might be the case. Just for curiosity's sake, I might try to drop the table
in-between the runs to see what happens.
JP
-Original Message-
From: sqlite-users-boun...@sqlite.org
On Jun 16, 2009, at 2:53 PM, Kalyani Phadke wrote:
> I am using sqlite version 3.6.15.I upgraded my database from sqlite
> 3.5.4 to 3.6.15. I am not able to open database from sqlite3.exe cmd
> utility .
> I can do sqlite3.exe test.DB but if type
> .database ,it gives me error malformed database
Ok, but you might have to run this a bit longer (or more iterations). My first
10 runs give the following results (in seconds):
1 15,681
2 16,010
3 16,093
4 16,168
5 16,057
6 16,585
7 16,114
8 16,596
9 16,115
10 16,270
Jumping around a
I am using sqlite version 3.6.15.I upgraded my database from sqlite
3.5.4 to 3.6.15. I am not able to open database from sqlite3.exe cmd
utility .
I can do sqlite3.exe test.DB but if type
.database ,it gives me error malformed database schema.
pragma integrity_check , gives error malformed
Closing and opening again did not speed up steps 1-4, it actually slowed things
down even more. The curve from the beginning is a bit similar to a slightly
flattened log curve. When I closed the database and started the test again, a
similar curve appeared again, but now starting from where the
Jeffrey --
Maybe so, but I should have looked further in Jen's
reply at the built-in optimizations for DELETE on a
Table without any filters and not having TRIGGERS ...
sorry :(
-- kjh
On 06/16/2009 12:44 PM, Hoover, Jeffrey wrote:
> Wouldn't a period VACUUMing of the database alleviate
Wouldn't a period VACUUMing of the database alleviate fragmentation?
- Jeff
-Original Message-
From: sqlite-users-boun...@sqlite.org [mailto:sqlite-users-boun...@sqlite.org]
On Behalf Of Konrad J Hambrick
Sent: Tuesday, June 16, 2009 1:43 PM
To: General Discussion of SQLite Database
Jens --
I have seen Index and Record Fragmentation cause
the kind of slowdowns you are describing.
Especially when there are a lot deletes followed
by a 'largish' (100,000-record) batch of inserts.
I have found on occasion that the speed is more
consistent if I can find a way to drop and
On 16 Jun 2009, at 4:46pm, Swithun Crowe wrote:
> How about having an extra column for each column that you want to
> search
> in? In the extra column, have a plain lowercase ASCII version of the
> word.
> So, for 'Sào Paulo', have 'sao paulo'. You would need to write a small
> program to
Fred Williams
wrote:
> What! The standard is lacking? Will wonders never cease?
>
> Thanks, you saved me from some boring reference reading.
>
> Still think, logically speaking, the construct should throw an error
> message, rather than make stealth changes to the
On 16 Jun 2009, at 5:32pm, Fred Williams wrote:
> Still think, logically speaking, the construct should throw an error
> message, rather than make stealth changes to the expected result.
I suppose it might be possible to provide SQLite with a 'conform to
standard' mode you could select using
On 16 Jun 2009, at 4:31pm, Konrad J Hambrick wrote:
> Is there a way to replace step 5 (delete everything from the table)
> with a System-Level rm / unlink and a sqlite CREATE TABLE ?
JP can totally do that, but his original question was why the system
gradually gets slower and slower. and
On Tue, Jun 16, 2009 at 10:57 AM, Jens Páll
Hafsteinsson wrote:
> The key factor here is not the total time taken to perform these operations
> but the fact that the time is *increasing* for each run. I am looking for
> consistency in that I need to be able to let the
On 16 Jun 2009, at 1:06pm, Jens Páll Hafsteinsson wrote:
> 1. start a transaction
>
> 2. insert 1000 records
>
> 3. commit
>
> 4. repeat steps 1-3 100 times
>
> 5. delete everything from the table
>
> 6. Start again at step 1
Okay, so do that until it's slow,
What! The standard is lacking? Will wonders never cease?
Thanks, you saved me from some boring reference reading.
Still think, logically speaking, the construct should throw an error
message, rather than make stealth changes to the expected result.
Fred
-Original Message-
From:
Is this a violation of SQL Standards? Seems like it should be.
I'm not for dumbing down the system to compensate for the occasional idiot
programmer. After all, we are not the government, nor should we strive to
act like it (IMHO).
If someone chooses to attempt to create a
The key factor here is not the total time taken to perform these operations but
the fact that the time is *increasing* for each run. I am looking for
consistency in that I need to be able to let the application perform these
steps in constant time over a long period of time (months).
Dropping
Fred Williams
wrote:
> Is this a violation of SQL Standards? Seems like it should be.
As far as I can tell, CREATE TABLE AS SELECT is not part of SQL92
standard at all. If you want to work within the standard, don't use this
construct in the first place.
Igor
Thank you Pavel for the lesson. I learned to never underestimate the power of
self join.
Works :)
--- On Tue, 6/16/09, Pavel Ivanov wrote:
> From: Pavel Ivanov
> Subject: Re: [sqlite] Rather SQL quiestion
> To: "General Discussion of SQLite
Christophe Leske wrote:
>> You write your own comparison function that would consider these two
>> strings equal. See sqlite3_create_function, sqlite3_create_collation.
>>
> this problem pertains not only to Zürich, but to 24000 other entries,
> so
> I guess that this is no
Thanks for the pointers Kees. I'll keep those in mind when I go into trying to
optimize the database.
I was clearly way off track when I said I was perfectly aware of the issues
concerning the database. I completely forgot about taking the spin factor into
account as you and Jim have pointed
Is this a violation of SQL Standards? Seems like it should be.
I'm not for dumbing down the system to compensate for the occasional idiot
programmer. After all, we are not the government, nor should we strive to
act like it (IMHO).
If someone chooses to attempt to create a
Ah, of course. Thanks for the explanation Jim.
JP
-Original Message-
From: sqlite-users-boun...@sqlite.org [mailto:sqlite-users-boun...@sqlite.org]
On Behalf Of Jim Wilcoxson
Sent: 16. júní 2009 15:07
To: General Discussion of SQLite Database
Subject: Re: [sqlite] Database inserts
Try this:
select c.*
from players a, registrations b, players c
where a.device_id = b.device_id
and b.mesg_token='123456'
and a.table_group_id = c.table_group_id
and a.table_id = c.table_id
Pavel
On Tue, Jun 16, 2009 at 11:25 AM, Boris Ioffe wrote:
>
> Hello
Hmmm ...
Is there a way to replace step 5 (delete everything from the table)
with a System-Level rm / unlink and a sqlite CREATE TABLE ?
-- kjh
On 06/16/2009 10:06 AM, Jim Wilcoxson wrote:
> You are doing transactions here, which is a very different thing from
> normal disk I/O. Your CPU is
Hello folks,
I am having trouble understanding one very important concept about IN operator.
Let me show you my situation below.
sqlite> select sqlite_version();
sqlite_version()
3.3.6
*
Here is my table schema
CREATE
> You write your own comparison function that would consider these two
> strings equal. See sqlite3_create_function, sqlite3_create_collation.
>
Well,
this problem pertains not only to Zürich, but to 24000 other entries, so
I guess that this is no option for me.
And again, I am using the
On Tue, 16 Jun 2009 14:23:47 +, Jens Páll Hafsteinsson
wrote:
> Yes, I'm perfectly aware of this and hence I
> would expect the disk to be sweating like hell
> running this test while the CPU is relatively
> relaxed (given that sqlite is disk bound in
> this case and not
Christophe Leske wrote:
> - how can SQlite be instructed to return search results which include
> a
> special character in it?
> E.g. you search literally for "Zurich" on an englisch system and
> expect
> "Zürich" to be in the result set.
You write your own comparison
You are doing transactions here, which is a very different thing from
normal disk I/O. Your CPU is idle because it is waiting for the disk.
Your disk is idle because it is waiting for the platters to rotate
around again. The best case you can achieve on a 7200RPM disk is 120
transactions
>
>
> Ok, I can see this. But I cannot forsee the real implications
> though. If I
> have a statement like
>
> create table mytable(pnumber integer, name varchar(20), primary
> key(pnumber))
>
> what are the reported datatypes? still integer and varchar(20)?
Yes. But if you do
CREATE
On Jun 16, 2009, at 8:47 AM, A Drent wrote:
>
> From the docs I read that for the new version:
>
> a.. When new tables are created using CREATE TABLE ... AS SELECT ...
> the
> datatype of the columns is the simplified SQLite datatype (TEXT,
> INT, REAL,
> NUMERIC, or BLOB) instead of a copy of the
Yes, I'm perfectly aware of this and hence I would expect the disk to be
sweating like hell running this test while the CPU is relatively relaxed (given
that sqlite is disk bound in this case and not CPU bound).
But this is not happening; neither the disk nor the CPU are practically doing
Hi there,
i have written an application which runs under german and englisch
versions of Windows.
It includes a city databases which is ought to be searchable, yet there
are a couple of issues which are of more logical nature...
My shell application surrounding the sqlite database only
Remember the implications of Moore's law and how much time has passed.
CPU speed is much faster than memory speed.
Memory speed is much faster than disk access.
This is why hardware folks play all sorts of tricks with pipelines, caches,
interleaving, and parallelism.
For a single process that
In step 5 I execute "delete from t1" without any where clause.
I haven't monitored the disk space used (does sqlite use temporary files beside
the database file?) but the database file itself has been fixed in size at
around 12MB (12.461.056 bytes) the whole time.
The load on the disk is
No, but I downloaded 3.6.15 and am running it through the test now. Will let
you know the results.
JP
-Original Message-
From: sqlite-users-boun...@sqlite.org [mailto:sqlite-users-boun...@sqlite.org]
On Behalf Of hiral
Sent: 16. júní 2009 12:10
To: General Discussion of SQLite Database
On Jun 16, 2009, at 8:47 AM, A Drent wrote:
>
> From the docs I read that for the new version:
>
> a.. When new tables are created using CREATE TABLE ... AS SELECT ...
> the
> datatype of the columns is the simplified SQLite datatype (TEXT,
> INT, REAL,
> NUMERIC, or BLOB) instead of a copy
On 16/06/2009 10:47 PM, A Drent wrote:
> Sorry, something went wrong on the previous post.
*AND* on this one; you are starting a new topic but you included about
900 lines from today's digest!!
>>From the docs I read that for the new version:
>
> a.. When new tables are created using CREATE
A Drent wrote:
> a.. When new tables are created using CREATE TABLE ... AS SELECT ...
> the datatype of the columns is the simplified SQLite datatype (TEXT,
> INT, REAL, NUMERIC, or BLOB) instead of a copy of the original
> datatype from the source table.
>
> I don't know why this has been done
Sorry, something went wrong on the previous post.
>From the docs I read that for the new version:
a.. When new tables are created using CREATE TABLE ... AS SELECT ... the
datatype of the columns is the simplified SQLite datatype (TEXT, INT, REAL,
NUMERIC, or BLOB) instead of a copy of the
>From the new version I read:
- Original Message -
From:
To:
Sent: Tuesday, June 16, 2009 2:00 PM
Subject: sqlite-users Digest, Vol 18, Issue 62
> Send sqlite-users mailing list submissions to
> sqlite-users@sqlite.org
>
> To
dave lilley wrote:
> Many thanks to all who have replied,
>
> I know understand the difference and shall use that approach to creating my
> queries.
>
> regarding the "ruby way" it was more how I saw saving code typing by
> injection different table, field and user data into one query thus saving
How do you do step 5? Like "delete from table" or "delete from table
where ..."? Do you see any degrade in disk space used by database
along with slowness?
Pavel
On Tue, Jun 16, 2009 at 8:06 AM, Jens Páll
Hafsteinsson wrote:
> Hi
>
> I've been running some tests against sqlite
Sorry I got it.
Did you tried with 3.6.14.2.
Thank you.
-Hiral
On Tue, Jun 16, 2009 at 5:39 PM, hiral wrote:
> Hi,
>
> Can you please let me know which version of sqlite, are you using?
>
> Thank you.
> -Hiral
>
> On Tue, Jun 16, 2009 at 5:36 PM, Jens Páll
Hi,
Can you please let me know which version of sqlite, are you using?
Thank you.
-Hiral
On Tue, Jun 16, 2009 at 5:36 PM, Jens Páll Hafsteinsson
wrote:
> Hi
>
> I've been running some tests against sqlite and have found that inserts are
> gradually slowing down. Since I'm
Hi
I've been running some tests against sqlite and have found that inserts are
gradually slowing down. Since I'm new to sqlite I might be doing something
wrong and thought I'd ask here if anyone has seen this before or know what
might be causing this.
The test I'm running creates a database
Hi,
I am using sqlite-3.5.9 and observing a 'disk image malformed' error
nfs, on doing 'PRAGMA integrity_check' I got following messages...
SQLite version 3.5.9
Enter ".help" for instructions
sqlite> PRAGMA integrity_check;
*** in database main ***
Page 5275 is never used
wrong # of entries in
I have created Tables A & AS_FTS
"create table A (id integer primary key, string Text);"
"create virtual table AS_FTS (Name);"
and a trigger
"insert into A_FTS (rowid,Name) values (New.%@,New.%@);"
(and a not shown delete trigger).
Now, I enter two strings into tue table: one and two. I
Many thanks to all who have replied,
I know understand the difference and shall use that approach to creating my
queries.
regarding the "ruby way" it was more how I saw saving code typing by
injection different table, field and user data into one query thus saving
typing.
BUT in the interests
80 matches
Mail list logo