Thank you to Igor and Richard. I've studied this issue more, and still don't
have an answer, although I've not been able to reproduce it either. I'm not
using shared cache, and even if I did leave a database connection open, which
seems impossible since sqlite3_open(), sqlite3_finalize() and
Without additional information, SQLite guesses that the data_idx index will
narrow down the search to about 7 entries in the table. This is, of
course, a guess, but it is a reasonable guess for most indices. The
primary key, even though it is unique, has an IN clause with 50 entries, it
SQLite
On Fri, Jun 7, 2013 at 12:56 PM, Brad House wrote:
> I would expect all queries which specify the primary key components
>>> in the WHERE clause to use the Primary Key in the query plan, regardless
>>> of if ANALYZE has been run or not.
>>>
>>>
>> SQLite examines many
On 6/6/2013 21:56, Philip Goetz wrote:
I think the problem is that the Cygwin distribution has the wrong
version of SQLite, one built for unix.
Nope. And even if true, it wouldn't be the right explanation.
There are two major ways to build SQLite on Cygwin:
1. By default, building SQLite
On 06/07/2013 12:46 PM, Simon Slavin wrote:
On 7 Jun 2013, at 5:37pm, Brad House wrote:
I've modified my code to run an Analyze on startup to work around this,
but it obviously takes time to run and slows down startup.
I can't answer your question about why this happens
I would expect all queries which specify the primary key components
in the WHERE clause to use the Primary Key in the query plan, regardless
of if ANALYZE has been run or not.
SQLite examines many different strategies for evaluating each query. For
each strategy it tries to estimate the total
On 7 Jun 2013, at 5:37pm, Brad House wrote:
> I've modified my code to run an Analyze on startup to work around this,
> but it obviously takes time to run and slows down startup.
I can't answer your question about why this happens in the first place, but I
can tell you that
On Fri, Jun 7, 2013 at 12:37 PM, Brad House wrote:
> I would expect all queries which specify the primary key components
> in the WHERE clause to use the Primary Key in the query plan, regardless
> of if ANALYZE has been run or not.
>
SQLite examines many different strategies
I would expect all queries which specify the primary key components
in the WHERE clause to use the Primary Key in the query plan, regardless
of if ANALYZE has been run or not.
I would also think it would assume any index which covers the most
where-clause components would be the most efficient
On Fri, Jun 7, 2013 at 11:51 AM, Ryan Johnson
wrote:
> On 06/06/2013 10:52 AM, Gabriel Corneanu wrote:
>
>> In my opinion, count(*) is the same as count(rowid) (I see that even
>> count() is accepted); I could say it's even the same as count(x) (any
>> other
>>
On 06/06/2013 10:52 AM, Gabriel Corneanu wrote:
In my opinion, count(*) is the same as count(rowid) (I see that even
count() is accepted); I could say it's even the same as count(x) (any other
field).
Not quite... count(x) only counts rows having non-NULL x. Granted,
that's not a problem for
Quoth Drake Wilson , on 2013-06-07 08:18:05 -0500:
> Actually, I dropped a paragraph on the floor, sorry. It's probably better to
> use
> xRead for this, since in that case SQLite will manage its own memory for the
> cache
> of decrypted pages. The loss in that case, if
- Original Message -
From: Richard Hipp
To: phi...@blastbay.com ; General Discussion of SQLite Database
Sent: Friday, June 07, 2013 1:14 PM
Subject: Re: [sqlite] regexp.c
On Fri, Jun 7, 2013 at 5:30 AM, Philip Bennefall wrote:
Hello all,
Quoth Drake Wilson , on 2013-06-07 08:14:27 -0500:
> If you really want, you might be able to implement xFetch to allocate a shadow
> buffer, decrypt from the map into that, and return that pointer. Since it's
> designed for accessing maps directly, though, I don't see it
Quoth Paul Vercellotti , on 2013-06-07 01:07:35 -0700:
> The current system does in-place decryption on read (no copying), which adds
> almost no overhead to the operation. Now in-place decryption with a
> memory-mapped file seems like it would dirty the page, which could
>
On 6 Jun 2013, at 2:30pm, Maxime Gerum wrote:
> but in c++, if i make request[4]= ”.mode csv”
Commands starting with "." are specially built into the SQLite shell tool
application. They are not part of SQLite itself and cannot be used just by
using API calls. Sorry.
Is this a repeated post?
You cannot use shell commands as Sql statements in c api. Write your own export
routine. Here is one I wrote.
https://github.com/fnoyanisi/sqlite3_capi_extensions
On 06/06/2013, at 11:02 PM, Maxime Gerum wrote:
>
> Hello, i’m french, sorry if
Hello, i’m french, sorry if i make errors of langage In my school project, i
use sqlite3 with c++, i can make any request such as create table or select
*from, no problems. But i would export my database in a csv file, in sqlite3>
.mode csv
Hello, i’m french, sorry if i make errors of langage In my school project, i
use sqlite3 with c++, i can make any request such as create table or select
*from, no problems. But i would export my database in a csv file, in sqlite3>
.mode csv
On Fri, 07 Jun 2013 13:12:14 +0200
Clemens Ladisch wrote:
> Eduardo Morras wrote:
> > where t.a = NULL
>
> where t.a IS NULL
>
> (NULL compares as not equal to any value, including itself.)
OPppss you're right. Thought too fast and wrote even faster :(
>
> Regards,
>
On Fri, Jun 7, 2013 at 5:30 AM, Philip Bennefall wrote:
> Hello all,
>
> I was having a quick look at the extensions provided in ext/misc in the
> source tree, and I am pleasantly surprised at the number of useful things
> that I found in there that I didn't know about. I was
Eduardo Morras wrote:
> where t.a = NULL
where t.a IS NULL
(NULL compares as not equal to any value, including itself.)
Regards,
Clemens
___
sqlite-users mailing list
sqlite-users@sqlite.org
On Thu, 6 Jun 2013 10:53:55 -0400
Richard Hipp wrote:
> On Thu, Jun 6, 2013 at 10:52 AM, Gabriel Corneanu > wrote:
>
> > Strange is, count(*) uses the cover index for a but "select count(a)" does
> > NOT use the same cover index...
> >
>
> count(a)
This is not at all my case ...
I don't obviously write 1 by 1, but using blocks of data ( array of
struct ), virtual tables wrappers, and "insert ... select".
This way I can achieve >200k rec/s, or at least 100k when having some
more fields.
Right now I'm completely CPU bound, it's 100% load
On Thu, 6 Jun 2013 08:15:57 -0400
Richard Hipp wrote:
> On Thu, Jun 6, 2013 at 8:05 AM, Eduardo Morras wrote:
>
> >
> > Hi,
> >
> > Is there an official list of assigned application id sqlite header? If
> > exist, How can I register my application-id?
> >
>
Hello all,
I was having a quick look at the extensions provided in ext/misc in the
source tree, and I am pleasantly surprised at the number of useful things
that I found in there that I didn't know about. I was wondering about the
regular expression parser in particular. Does it support
On 7 Jun 2013, at 8:57am, Gabriel Corneanu wrote:
> BTW I found this by opening some file over network, which of course made
> everything worse.
> [...]
> Not that I really need, but I have to support specified data rates up to 100k
> records / second.
Maximum
Hi there,
We use a vfs-based obfuscation system using the old xRead and xWrite methods.
I'm wondering if these can be adapted to work with the new memory-mapped i/o
functionality in a way that still has the advantages of memory-mapped i/o?
The current system does in-place decryption on read
I admit I didn't think (or didn't even read in detail) about technical
implementation.
This is an extract from analyzer:
*** Table AE_DATA
Percentage of total database.. 99.89%
Number of entries. 1030371
Bytes of
29 matches
Mail list logo