Re: [sqlite] Is something special to load fts1.dll in Tcl?

[Gerry Snyder]

[EMAIL PROTECTED] wrote:

FTS1 is built into the starkit.  So if you use the starkit, there
is nothing to do to enable full-text search.  It's just there.

I thought I did the same for the TCL DLL as well - FTS1 is built in.
But I admit that I might have not gotten that exactly right.  Did
you try just using the FTS1 extension without loading it first?
  

No, (blush), I did not. Should have, but did not think of it.

It works just as advertised, right out of the box.

Thanks much.

Gerry

-
To unsubscribe, send email to [EMAIL PROTECTED]
-

Re: [sqlite] Is something special to load fts1.dll in Tcl?

[drh]

Gerry Snyder <[EMAIL PROTECTED]> wrote:
> I am eager to try out full text search using SQLite and Tcl on WinXP.
> 
> I downloaded the Tcl bindings and executable for SQLite 3.3.8 and the 
> fts1.dll file, and copied them all (after unzipping) to a directory. The 
> fts1.dll loads fine using the stand-alone executable, but I get a "not 
> authorized" error using Tcl (both the latest freewrap and an 8.5a4 kit).
> 

FTS1 is built into the starkit.  So if you use the starkit, there
is nothing to do to enable full-text search.  It's just there.

I thought I did the same for the TCL DLL as well - FTS1 is built in.
But I admit that I might have not gotten that exactly right.  Did
you try just using the FTS1 extension without loading it first?
--
D. Richard Hipp  <[EMAIL PROTECTED]>


-
To unsubscribe, send email to [EMAIL PROTECTED]
-

Re: [sqlite] new sqlite-based webserver

[John Stanton]

Martin Jenkins wrote:

Fred Williams wrote:


What'ch got it running on, a 286?  Tried three different times and
got tired waiting all three times :-(



Seems ok now. Quite fast even.

mj

- 


To unsubscribe, send email to [EMAIL PROTECTED]
- 



It is still dead.  Address 62.194.135.186 doesn't respond.  Traceroute 
fails.


-
To unsubscribe, send email to [EMAIL PROTECTED]
-

[sqlite] Is something special to load fts1.dll in Tcl?

[Gerry Snyder]

I am eager to try out full text search using SQLite and Tcl on WinXP.

I downloaded the Tcl bindings and executable for SQLite 3.3.8 and the 
fts1.dll file, and copied them all (after unzipping) to a directory. The 
fts1.dll loads fine using the stand-alone executable, but I get a "not 
authorized" error using Tcl (both the latest freewrap and an 8.5a4 kit).


Any clues as to what I'm doing wrong? (The missing numbers below are 
from typos.)


TIA, Gerry

(fulltextsearch) 1 % load tclsqlite3.dll

(fulltextsearch) 3 % sqlite3 sq gigo

(fulltextsearch) 6 % sq eval "select load_extension('fts1.dll')"
not authorized
(fulltextsearch) 7 % sq eval "select load_extension('libfts1.dll')"
not authorized
(fulltextsearch) 8 %



-
To unsubscribe, send email to [EMAIL PROTECTED]
-

Re: [sqlite] new sqlite-based webserver

[Martin Jenkins]

Fred Williams wrote:

What'ch got it running on, a 286?  Tried three different times and
got tired waiting all three times :-(


Seems ok now. Quite fast even.

mj

-
To unsubscribe, send email to [EMAIL PROTECTED]
-

[sqlite] creating a view on an attached database

[P Kishor]

I am attaching external database 'b' to database 'a',

ATTACH DATABASE b AS d_b;

and then trying to create a view that uses tables from the attached database,

CREATE VIEW v AS
 SELECT d_b_t.col, d_a_t.col
 FROM d_b.tbl AS d_b_t JOIN tbl AS d_a_t ON...

I get the following error --

view v cannot reference objects in database d_b(1)

is that not allowed?

--
Puneet Kishor http://punkish.eidesis.org/
Nelson Inst. for Env. Studies, UW-Madison http://www.nelson.wisc.edu/
Open Source Geospatial Foundation https://edu.osgeo.org/

-
To unsubscribe, send email to [EMAIL PROTECTED]
-

Re: [sqlite] Re: SQLite Order BY

[John Stanton]

A. Pagaltzis wrote:

* John Stanton <[EMAIL PROTECTED]> [2006-10-09 19:35]:


Sorting data is time consuming, a physical law is involved. At
best it is an nlog(n) process.



Only when you sort by comparing elements with each other. Bucket
sort runs in O(n), f.ex. And quantum sort is O(1). ;-) Algorithms
that run faster than O(n log n) are very rarely practical,
however.

Regards,
Congratulations, you have your next project, modifying Sqlite to use a 
sort performing better than nlog(n).  We look forward to it and wish you 
well.


-
To unsubscribe, send email to [EMAIL PROTECTED]
-

[sqlite] Re: SQLite Order BY

[A. Pagaltzis]

* John Stanton <[EMAIL PROTECTED]> [2006-10-09 19:35]:
> Sorting data is time consuming, a physical law is involved. At
> best it is an nlog(n) process.

Only when you sort by comparing elements with each other. Bucket
sort runs in O(n), f.ex. And quantum sort is O(1). ;-) Algorithms
that run faster than O(n log n) are very rarely practical,
however.

Regards,
-- 
Aristotle Pagaltzis // 

-
To unsubscribe, send email to [EMAIL PROTECTED]
-

Re: [sqlite] new sqlite-based webserver

[John Stanton]

Gunter,
I cannot access your web site.  The address is resolved, then the server 
hangs.  It would be interesting to look at your www server.


Günter Greschenz wrote:

hi,

i dont know if anyone is interested in my new open source project...
i implemented a little webserver with javascript as backend-language
(1.7 from mozilla 2.0.rc1) and sqlite as datastorage (3.3.7)
the server is serving himself on "http://greschenz.dyndns.org; :-)
the source can be downloaded at "http://greschenz.dyndns.org/download/gas/;
its compilable with gcc(linux) and vs2005(win32)

please send me any comments / ideas...

cu, gg



- 


To unsubscribe, send email to [EMAIL PROTECTED]
- 






-
To unsubscribe, send email to [EMAIL PROTECTED]
-

RE: [sqlite] new sqlite-based webserver

[Fred Williams]

What'ch got it running on, a 286?  Tried three different times and got
tired waiting all three times :-(

Not going to stir much interest with response times like that!


> -Original Message-
> From: Günter Greschenz [mailto:[EMAIL PROTECTED]
> Sent: Monday, October 09, 2006 12:04 PM
> To: sqlite-users@sqlite.org
> Subject: [sqlite] new sqlite-based webserver
>
>
> hi,
>
> i dont know if anyone is interested in my new open source project...
> i implemented a little webserver with javascript as backend-language
> (1.7 from mozilla 2.0.rc1) and sqlite as datastorage (3.3.7)
> the server is serving himself on "http://greschenz.dyndns.org; :-)
> the source can be downloaded at
> "http://greschenz.dyndns.org/download/gas/;
> its compilable with gcc(linux) and vs2005(win32)
>
> please send me any comments / ideas...
>
> cu, gg
>
>
>
> --
> ---
> To unsubscribe, send email to [EMAIL PROTECTED]
> --
> ---
>


-
To unsubscribe, send email to [EMAIL PROTECTED]
-

Re: [sqlite] SQLite Order BY

[John Stanton]

Sorting data is time consuming, a physical law is involved.  At best it 
is an nlog(n) process.  Sorting a result set or sorting the keys to 
build an index invokes the nlog(n) timing.


If you cannot tolerate time spent sorting, you need to redesign your 
database to avoid sorts, or to maintain reference data in sorted 
sequence via an index and use that sequence repeatedly.


This is all IT101, nothing to do with Sqlite.

Manzoor Ilahi wrote:

Thanks shiming
we have tried this method as u suggested , but creating an index on a 5 
million records

take a long time, nearly 4 seconds.

And in our case the Database contain 5 Million to 30 or 50 million 
records. so if I try to create additional indices on these tables it 
takes too much time.


any other suggestions if u have please.
Best regards,
Manzoor Ilahi

//-- 



You can try creating two additional indexes, one on hvh5m.Column4 and 
the other on hvh5m.Column8. They should speed things up.


Best regards,
He Shiming

From: "He Shiming" <[EMAIL PROTECTED]>
Reply-To: sqlite-users@sqlite.org
To: 
Subject: Re: [sqlite] SQLite Order BY Date: Mon, 9 Oct 2006 11:59:46 
+0800


//-- 




Dar All,

Can someone guide me, why the ORDER BY in SQLite is too slow. we 
observed that if the   change in data values is small then ORDER BY 
works better . but if the change is big then the performance is very 
slow. Even sometime I can not finish a query execution, and wait and 
may be sometime need to kill the process.


also another point that i observed is ,  when ..order by fieldName , 
if fieldName  is varchar,the excutiion is faster ,and if fieldName is 
float,the execution can't be finished!



Any Idea Please

thanks

Manzoor Ilahi

Queries that we tested and the Schema is given as under.


//-- 


select * from hvh5m,itm where hvh5m.Column4=itm.Icol1 order by Column8

this SQl can't be finished

//-- 



select * from hvh5m,itm where hvh5m.column4=itm.Icol1 order by 
itm.Icol1;


this sql can be finished!

//-- 



hvh5m: column1 varchar(8 );
column2 varchar(4);
Column3 varchar(2);
Column4 varchar(4)£»
column5 varchar(7);
column6 varchar(8 );
column7 varchar(1);
Column8float(10£¬2);

itm£º Icol1 varchar(4);
itemname varchar(20);
Icol3 varchar(1);
Icol4 varchar(1);
Icol5 varchar(1);
Icol6 varchar(1);
Icol7 varchar(1);
Icol8 varchar(1);
Icol 9 varchar(1);
Icol 10 varchar(1);
Icol 11 varchar(1);
Icol 12 varchar(1);
Icol 13 varchar(1);
Icol 14 varchar(1);
Icol 15 varchar(1);
Icol 16 varchar(1);

There is one index on itm.Icol1.



You can try creating two additional indexes, one on hvh5m.Column4 and 
the other on hvh5m.Column8. They should speed things up.


Best regards,
He Shiming


- 


To unsubscribe, send email to [EMAIL PROTECTED]
- 





_
Be seen and heard with Windows Live Messenger and Microsoft LifeCams 
http://clk.atdmt.com/MSN/go/msnnkwme002001msn/direct/01/?href=http://www.microsoft.com/hardware/digitalcommunication/default.mspx?locale=en-us=hmtagline 




- 


To unsubscribe, send email to [EMAIL PROTECTED]
- 






-
To unsubscribe, send email to [EMAIL PROTECTED]
-

Re: [sqlite] results of SELECT contains carriage returns

[P Kishor]

On 10/9/06, T <[EMAIL PROTECTED]> wrote:

Hi He,

>> sqlite3 Disney.db "SELECT * FROM Characters"
>>
>> When I get the result of a SELECT statement, the output has
>> columns separated by pipe characters and rows by new lines. So, if
>> a value contains a return, it prematurely starts a new line, and
>> messes up my output result.

> Well, you didn't say which midware or platform were you using. I'll
> talk in the perspective of C API.

At the moment I'm just prototyping, on a Mac, using Xcode for the UI
and AppleScript calling sqlite via shell commands. I may move
sections of it over to C in the long term, once I have it all
working. But for now, my only interface to sqlite is via the sqlite3
shell command.

> In reality, we never parse query results directly from the "table"
> output from command line shell.



..


What is QuickLite?

QuickLite is a Cocoa wrapper for SQLite, a C library that implements
an embeddable SQL database engine.

See more at 

--
Puneet Kishor http://punkish.eidesis.org/
Nelson Inst. for Env. Studies, UW-Madison http://www.nelson.wisc.edu/
Open Source Geospatial Foundation https://edu.osgeo.org/

-
To unsubscribe, send email to [EMAIL PROTECTED]
-

[sqlite] [ANN] New DBManager Professional 3.4.0 Enterprise Edition

[COS]

DBTools Software is pleased to announce the new DBManager Professional 3.4.0
Enterprise Edition.
This is a major update with plenty of bugs fixed and new improvements to the
application. See the details:

NEW FEATURES AND IMPROVEMENTS

- Added Context Menu for Default Servers
- Diagram Designer (added option [Save] to table menu)
- Diagram Designer (added [*] mark when table is new)
- Implemented AutoConfigureServers on first run
- Objects are listed inside Namespace/Schema (If supported)
- Better support for SQlite Unicode databases
- ODBC now lists schemas/databases (if supported by the driver)

FIXES

- Fixed View Selected and Double Clicked in Objects Bar generated error
- Fixed SQL Designer (Failed to added [View])
- Fixed Oracle not showing View Columns
- Fixed Firebird/Interbase not have autoinc on table designer
- Diagram Designer (Fixed crash when dragging fields and changing name)
- Fixed Diagram Designer crash when create new table and drag column
with new name
- Procedure Builder accumulating error messages on compilation
- Fixed crash when connecting to Workgroup when no catalog is opened
- Fixed (Database missing when saving attached servers)
- Fixed Syntax error for INSERT MSAccess when column is image
- Fixed MySQL CreateIndex (missing columns quote)
- Fixed SQL Server Table copy did not SET IDENTITY_INSERT for inserts
- Fixed Diagram not asking to save before closing window
- Fixed Server Manager didn't show default database
- Fixed several memory leaks
- Fixed Server Manager Loosing some parameters when editing an existent
connection
- Fixed Server Manager loosing typed group when inserting new ODBC
connection
- Check ODBC Datatypes available when connecting (removed duplicates)
- Fixed Datasheet View now saves its layout
- Fixed Table Properties correctly shows foreign keys (PKTable was
empty/schema)
- Fixed ODBC not counting table rows when required
- Fixed MSSQL missing quotes when creating constraints
- Fixed Form did not show image for the first record

The current version  is now available in our Downloads Center
(http://www.dbtools.com.br/EN/downloads).

WHAT IS DBMANAGER?

DBManager Professional is a Windows application for database management,
with built-in support for Oracle, MSSQL Server/MSDE, Sybase, MSAccess,
MySQL, PostgreSQL, Interbase, Firebird, DBF Tables and ODBC Datasources. It
has all set of features which can include:

- Object Management (Database, Tables, Indexes, Foreign Keys, Check
Constraints, Domains, Views, Stored Procedures, etc)
- Query Editor and Query Builder with planner and debugger capabilities
- Database and Server Monitoring charts
- Diagram, Report and Form Designers
- Database Documenter (Printable and in HTML Format)
- Objects Browser (shows objects dependencies)
- Task Builder (to create complex tasks)
- Wizards to Import and Export Data including: Database Migration, MSAccess,
MSExcel, ODBC, Text Files (Delimited, Fixed), HTML, XML, Database Comparer
to synchronize structure
- and much more

For a full list of features please check
http://www.dbtools.com.br/EN/dbmanagerpro.

For a limited period of time we are offering a special price in our product
list, which makes it possible to purchase the Master Pack (which supports
all database engines) with discounts of 81%. For more details see
http://www.dbtools.com.br/EN/order.php.

Best Regards,

DBTools Software
http://www.dbtools.com.br



-
To unsubscribe, send email to [EMAIL PROTECTED]
-

Re: [sqlite] results of SELECT contains carriage returns

[T]

Hi He,


sqlite3 Disney.db "SELECT * FROM Characters"

When I get the result of a SELECT statement, the output has  
columns separated by pipe characters and rows by new lines. So, if  
a value contains a return, it prematurely starts a new line, and  
messes up my output result.


Well, you didn't say which midware or platform were you using. I'll  
talk in the perspective of C API.


At the moment I'm just prototyping, on a Mac, using Xcode for the UI  
and AppleScript calling sqlite via shell commands. I may move  
sections of it over to C in the long term, once I have it all  
working. But for now, my only interface to sqlite is via the sqlite3  
shell command.


In reality, we never parse query results directly from the "table"  
output from command line shell.


My prototype is currently my reality ;-)

You are supposed to use a combination of sqlite3_open,  
sqlite3_prepare, sqlite3_step, which are the APIs provided by  
sqlite, to access the database. And use APIs like sqlite3_column_*  
to get the content of the query.


Thanks for the detail. I may have to take the plunge into C, but not  
yet. Unless someone has tips or a skeleton in how to integrate the  
SQLite C routines into Xcode.


Any other help appreciated.

Thanks,
Tom


-
To unsubscribe, send email to [EMAIL PROTECTED]
-

[sqlite] Re: any questions

[Igor Tandetnik]

mimuel1 <[EMAIL PROTECTED]> wrote:

- with access it is possible to integrate "pseudo-stored-procedure".
can I do this with SQLite too?


SQLite has triggers, which can be used as poor man's stored procedures. 
That's the only kind of "code" that can be stored in a database. Your 
program can install custom functions.



- where can I find ODBC-driver for SQLite?


Google is your friend: http://www.google.com/search?q=sqlite+odbc+driver

The first result is http://www.ch-werner.de/sqliteodbc/


- is it possible to split the SQLite-db-file? 4TB ist very big!


You can have a single connection access multiple SQLite database files 
with ATTACH command:

http://www.sqlite.org/lang_attach.html

Igor Tandetnik 



-
To unsubscribe, send email to [EMAIL PROTECTED]
-

[sqlite] Re: SQLite Order BY

[Igor Tandetnik]

He Shiming <[EMAIL PROTECTED]> wrote:

select * from hvh5m,itm where hvh5m.Column4=itm.Icol1 order by
Column8


You can try creating two additional indexes, one on hvh5m.Column4 and
the other on hvh5m.Column8. They should speed things up.


Actually, what would help here is a compound index on (Column4, Column8)

Igor Tandetnik 



-
To unsubscribe, send email to [EMAIL PROTECTED]
-

Re: [sqlite] transaction and tables

[drh]

chetana bhargav <[EMAIL PROTECTED]> wrote:
> Hi,
>
>   I have a query regarding transactions and tables.
>
>   Say in a DB if you have 4 tables, now in my code  I will say , "begin 
> transaction" and start modifying a table, say tbl1, now a journal would be 
> backed up for the tbl1 that I am going to modify, now if in the same 
> transaction if I try to modify a different table, say tbl2,  in the same DB, 
> will the multiple inserts into the tbl2 will be fast as we have already a 
> journal created. 

Yes.

>
>   If at this point the power goes off, will tbl2 be consistent?

Yes.  All updates are atomic.  Either both tbl1 and tbl2
will be updated or neither will.

>
>   And one more question,
>   If in my code if I call "begin transaction" on the same DB twice due to 
> some logical flow error, will it have any effect.
>

You will get an error on the second "begin transaction" but if
you just ignore that error you can continue normally and nothing
bad will come of it.
--
D. Richard Hipp  <[EMAIL PROTECTED]>


-
To unsubscribe, send email to [EMAIL PROTECTED]
-

[sqlite] transaction and tables

[chetana bhargav]

Hi,
   
  I have a query regarding transactions and tables.
   
  Say in a DB if you have 4 tables, now in my code  I will say , "begin 
transaction" and start modifying a table, say tbl1, now a journal would be 
backed up for the tbl1 that I am going to modify, now if in the same 
transaction if I try to modify a different table, say tbl2,  in the same DB, 
will the multiple inserts into the tbl2 will be fast as we have already a 
journal created. 
   
  If at this point the power goes off, will tbl2 be consistent?
   
  And one more question,
  If in my code if I call "begin transaction" on the same DB twice due to some 
logical flow error, will it have any effect.
   
   
  ...
  Chetana.


-
Get your email and more, right on the  new Yahoo.com 

[sqlite] any questions

[mimuel1]

Hi,
now I wrote a db-application with ms-access. Its only a single-user
application. But the size of access-dbs are limited.(2GB). Now I found
SQLite, limited at 4 TB.

My questions:
- with access it is possible to integrate "pseudo-stored-procedure". can I
do this with SQLite too?
- where can I find ODBC-driver for SQLite?
- is it possible to split the SQLite-db-file? 4TB ist very big!
- who knows links with the thema: "from access to SQLite"?

thanks michael 

-- 
View this message in context: 
http://www.nabble.com/any-questions-tf2409295.html#a6715543
Sent from the SQLite mailing list archive at Nabble.com.


-
To unsubscribe, send email to [EMAIL PROTECTED]
-

Re: [sqlite] Can anyone recommend some ISAM db to me?

[Paul Smith]

At 11:38 08/10/2006, you wrote:

Hi, all
After trying SQLite on my embedded platform, I feel that it's a 
little too complicated and time-consuming to my platform, especially 
the parsing.
So, could someone recommend several ISAM ones to me?(I'm a newbie of 
database*^_^*)


You could have a look at Codebase (www.codebase.com) - whether it 
will work 'out of the box' on your platform or not I can't say, but 
it comes with source code, and is royalty free, and is quick. We 
dumped it for SQLite because we wanted the flexibility of SQL, but 
otherwise it's a decent DB engine supporting xBASE database files.




PaulVPOP3 - Internet Email Server/Gateway
[EMAIL PROTECTED]  http://www.pscs.co.uk/



-
To unsubscribe, send email to [EMAIL PROTECTED]
-

Re: [sqlite] SQLite Order BY

[He Shiming]

Thanks shiming
we have tried this method as u suggested , but creating an index on a 5 
million records

take a long time, nearly 4 seconds.

And in our case the Database contain 5 Million to 30 or 50 million 
records. so if I try to create additional indices on these tables it takes 
too much time.


any other suggestions if u have please.
Best regards,
Manzoor Ilahi



I don't understand. Index creation is a one time deal, it's like tables 
(there are temporary indexes though, which are also similar to tables). You 
don't have to recreate indexes before queries. It only cost you 4 seconds, 
but future queries will be a lot faster. Newly inserted rows will be 
automatically indexed. Did you compare the performance before and after the 
creation of indexes? In my experience, indexes do improve performance a lot 
as long as you've indexed the correct columns, i.e. the columns appearing in 
where clause and order by clause.


Best regards,
He Shiming 



-
To unsubscribe, send email to [EMAIL PROTECTED]
-

[sqlite] error on sqlite3_close

[Kibae Shin]

Hi.

in sqlite3.3.8(latest)
 ./configure --prefix=/sw CFLAGS=-g CPPFLAGS=-g

1 thread : 1 db model

Error on many thread's request(about 1k/sec) to same db file.

This problem repeated sometimes.

I use sqlite3_commit_hook, sqlite3_rollback_hook, sqlite3_update_hook  
and sqlite3_set_authorizer APIs.


--- source code --
void client_safe_exit( struct thread_t *self )
{
db_close( >db );
}

int db_close( struct db_t *db )
{
int result;

if ( db->opened == TRUE )
{
db_err_free( db );
/*XXX ERROR */
result = sqlite3_close( db->db );
db->opened = FALSE;
db->db = NULL;
return result;
}
return SQLITE_OK;
}
--- gdb back trace result 

sqlitedbms(21573,0x1820600) malloc: ***  Deallocation of a pointer  
not malloced: 0x50a0e1; This could be a double free(), or free()  
called with the middle of an allocated block; Try setting environment  
variable MallocHelp to see tools to help debug


Program received signal EXC_BAD_ACCESS, Could not access memory.
Reason: KERN_INVALID_ADDRESS at address: 0x350c
[Switching to process 21573 thread 0x3f7f]
0x00226aba in sqlite3_close (db=0x521670) at ./src/main.c:155
155   pNext = pFunc->pNext;
(gdb) q
The program is running.  Exit anyway? (y or n) t^Hy^H^C
Please answer y or n.
The program is running.  Exit anyway? (y or n) n
Not confirmed.
(gdb) bt
#0  0x00226aba in sqlite3_close (db=0x521670) at ./src/main.c:155
#1  0x3996 in db_close (db=0x1811824) at libsqlite3.c:101
#2  0x56c2 in client_safe_exit (self=0x1811600) at proc.c:706
#3  0x5d12 in client (data=0x1811600) at proc.c:379
#4  0x90023d87 in _pthread_body ()
(gdb) bt full
#0  0x00226aba in sqlite3_close (db=0x521670) at ./src/main.c:155
pFunc = (FuncDef *) 0x3500
pNext = (FuncDef *) 0x3500
i = (HashElem *) 0x52aff0
j = 2
#1  0x3996 in db_close (db=0x1811824) at libsqlite3.c:101
result = 889192448
db = (struct db_t *) 0x1811824
#2  0x56c2 in client_safe_exit (self=0x1811600) at proc.c:706
self = (struct thread_t *) 0x1811600
#3  0x5d12 in client (data=0x1811600) at proc.c:379
pollfd = {
  fd = 16,
  events = 27,
  revents = 19
}
poll_res = 889192448
read_res = 889192448
auth_res = 889192448
db_res = 74352
result = 889192448
http_res = 889192448
data = (void *) 0x1811600
#4  0x90023d87 in _pthread_body ()
No symbol table info available.
(gdb)


-
To unsubscribe, send email to [EMAIL PROTECTED]
-

Re: [sqlite] SQLite Order BY

[Manzoor Ilahi]

Thanks shiming
we have tried this method as u suggested , but creating an index on a 5 
million records

take a long time, nearly 4 seconds.

And in our case the Database contain 5 Million to 30 or 50 million records. 
so if I try to create additional indices on these tables it takes too much 
time.


any other suggestions if u have please.
Best regards,
Manzoor Ilahi

//--
You can try creating two additional indexes, one on hvh5m.Column4 and the 
other on hvh5m.Column8. They should speed things up.


Best regards,
He Shiming

From: "He Shiming" <[EMAIL PROTECTED]>
Reply-To: sqlite-users@sqlite.org
To: 
Subject: Re: [sqlite] SQLite Order BY Date: Mon, 9 Oct 2006 11:59:46 +0800


//--

Dar All,

Can someone guide me, why the ORDER BY in SQLite is too slow. we observed 
that if the   change in data values is small then ORDER BY works better . 
but if the change is big then the performance is very slow. Even sometime 
I can not finish a query execution, and wait and may be sometime need to 
kill the process.


also another point that i observed is ,  when ..order by fieldName , if 
fieldName  is varchar,the excutiion is faster ,and if fieldName is 
float,the execution can't be finished!



Any Idea Please

thanks

Manzoor Ilahi

Queries that we tested and the Schema is given as under.


//--
select * from hvh5m,itm where hvh5m.Column4=itm.Icol1 order by Column8

this SQl can't be finished

//--

select * from hvh5m,itm where hvh5m.column4=itm.Icol1 order by itm.Icol1;

this sql can be finished!

//--

hvh5m: column1 varchar(8 );
column2 varchar(4);
Column3 varchar(2);
Column4 varchar(4)£»
column5 varchar(7);
column6 varchar(8 );
column7 varchar(1);
Column8float(10£¬2);

itm£º Icol1 varchar(4);
itemname varchar(20);
Icol3 varchar(1);
Icol4 varchar(1);
Icol5 varchar(1);
Icol6 varchar(1);
Icol7 varchar(1);
Icol8 varchar(1);
Icol 9 varchar(1);
Icol 10 varchar(1);
Icol 11 varchar(1);
Icol 12 varchar(1);
Icol 13 varchar(1);
Icol 14 varchar(1);
Icol 15 varchar(1);
Icol 16 varchar(1);

There is one index on itm.Icol1.



You can try creating two additional indexes, one on hvh5m.Column4 and the 
other on hvh5m.Column8. They should speed things up.


Best regards,
He Shiming


-
To unsubscribe, send email to [EMAIL PROTECTED]
-



_
Be seen and heard with Windows Live Messenger and Microsoft LifeCams 
http://clk.atdmt.com/MSN/go/msnnkwme002001msn/direct/01/?href=http://www.microsoft.com/hardware/digitalcommunication/default.mspx?locale=en-us=hmtagline



-
To unsubscribe, send email to [EMAIL PROTECTED]
-